For Immediate Release

Recent Episodes

SEP 28, 2026

FIR #529: Fake Quotes from Fake Experts Find Their Way Into Top-Tier Media

A Press-Gazette investigation found that dozens of experts quoted in the mainstream media don’t actually exist, and PR agencies are unwittingly sharing them with the press. This poses a credibility issue for the agencies and the media. It requires a layer of verification unlike anything we’ve faced before. Also in this episode, crisis communication has long been housed in the PR or communications department. Should it be a governance responsibility of the C-suite? That’s the case made in a new report from the Chartered Institute of Public Relations and the Institute of Directors. In his Tech Report, Dan York covers an initiative to close the AI language gap, including video moves by YouTube and Bluesky, smart glasses, and enhancements to Threads’ podcast toolkit. Links from this episode: Chartered Institute of Public Relations and Institute of Directors publish Crisis Communication Handbook for Boards Executive Reputation: Priced, Prized and Easily Lost Kate Winick: This Is a Bad Brand Clapback Converse Ad Sparks Controversy Over Racist Imagery Virtual Reality: The Widely Quoted Media Experts Who Are Not What They Seem Qwoted Bans AI-Generated Expert Profiles as PR Spam Floods Reporters’ Inboxes Prominent Art Therapist Quoted by Vice, Forbes and Others Is AI-Generated The Rise of Fake AI Experts: A Wake-Up Call for the Community The Rise of the Fake Expert: How to Remain Authentic in an Inbox of Fakes Links from Dan York’s Tech Report: Closing AI’s Language Gap: Why the Internet Society Is Joining a Global Commitment Bluesky Now Lets You Upload 10-Minute-Long Videos What’s New at Made on YouTube 2026? New YouTube Live Tools for Fan Funding, Dubbing and More Everything Announced at Meta Connect 2026 Snap Launches $2,195 Specs Smart Glasses Threads Expands Podcast Toolkit for Creators and Listeners The next monthly, long-form episode of FIR is tentatively scheduled to drop on Monday, October 26. We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected]. Special thanks to Jay Moonah for the opening and closing music. You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog. Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients. Raw Transcript: Shel Holtz: Hi, everybody, and welcome to episode 529 of For Immediate Release. I’m Shel Holtz in Concord, California. Neville Hobson: And I’m Neville Hobson in Somerset in the UK. Shel Holtz: Neville, it is wonderful to be back together with you again. Neville Hobson: It’s been a bit of a strange few weeks because of some health issues I’m working through. They’ve affected my availability to record our midweek shows. I was particularly keen to make sure we could record the topics we’re discussing today. Hopefully, by the time our next monthly episode comes around, I’ll be close to normal again. Shel Holtz: My fingers are certainly crossed. To accommodate all of this, we’re recording a shortened version of our monthly episode: two topics plus Dan York’s report. We’ll also have a quick recap of the two episodes since our last monthly show. Neville Hobson: Episode 527 was a particularly interesting long-form show with six topics. Our lead story was about a senior manager who discovered that AI was rejecting job applications within seconds of their arrival. One reason was that applicants had graduated from college before 2010, even though the instruction was to surface candidates who displayed “youthful energy.” What’s worse, the talent acquisition manager saw no liability issue because AI had made the decision. If I recall correctly, she said, “We’re okay. We have no liability because the AI made the decision.” If it weren’t so serious, it would be hilarious. We also discussed an expert witness who used ChatGPT to prepare a report for a legal case. The prompts he used—not just the AI’s output—were uncovered during discovery and entered into evidence. You have to check these things. Then, in episode 528, we talked about fandom. We explored the differences among fans, audiences, and communities. A fan’s identity can become tied to a team, artist, product, or other object of passion. It’s no surprise that marketers want their brands to inspire that kind of excitement, but not everything has the potential to develop a fan base. We received a couple of comments about that episode. Shel Holtz: Brian Kilgore in Toronto wrote, “Good show. In my life, I’ve worn Roy Rogers, Gene Autry, and Hopalong Cassidy clothes, an MG T-shirt, and a sweatshirt I designed using ASCII code as the hook for CNCP telecommunications. One cap that I’ve lost says ‘Nikon Most Excellent.’” Brian has been a fan of quite a few things. We also received an audio comment from our tech reporter, Dan York. Dan York: Hey, Shel and Neville. I had a comment on episode 528. I found myself cringing when I heard marketers talking about how to grow a fandom. You have fans, and marketers should find where those fans are and engage with them. They shouldn’t be trying to build a fandom. You were right that fandom can be part of someone’s identity. My neighbors across the street are big Mazda Miata owners. They each have one, and they go on rallies and tours with other owners. It’s a community and a group, but being Miata owners is also part of who they are. To its credit, the local Mazda dealership supports that community. There’s a lot of fandom on Reddit, too. I own a Hyundai Santa Cruz, a little vehicle that’s rather like an SUV with a truck bed on the back. Its subreddit has people sharing photos, asking questions, and talking about what they’re doing with their vehicles. Some are passionate about these little trucks. I’ve seen that with other vehicles as well. Marketers and communicators, please don’t talk about trying to create a fandom. Find the fans where they are. Shel Holtz: Thanks, Dan. I couldn’t agree more. If there’s a fan base around something you make or represent, by all means, see how you can support it and become more visible to it. Trying to build a fan base around something that doesn’t naturally inspire one can feel artificial. I hadn’t thought to look for those communities on Reddit. I have a Ford Mustang Mach-E, and as soon as we’re done recording, I’m going to look for a Mach-E subreddit. I have to believe there is one. I appreciate the comment. I also want to mention Circle of Fellows. Episode 133, part two of our conversation on AI and the leadership moment for the communication profession, should be available by the time you hear this. Adrian Cropley returned, and we were joined by Mary Hills, Cyrus Mavalwala, and Robin McCasland. Our next episode is Thursday, October 29, at 6 p.m. Eastern. That hour accommodates Zora Artis, who will be joining us for a discussion of connection and leadership. Three other Fellows will join us as well. You’ll find details on the FIR Podcast Network website. Neville Hobson: Let’s talk about crisis communication—or, more precisely, governance. When we talk about crisis communication, many of us picture what happens once a crisis is underway: the holding statement, the spokesperson, media response, frantic internal calls, whether the CEO should go public, and how quickly the organization says something. But a new handbook from the Chartered Institute of Public Relations’ Crisis Communications Network, published with the UK’s Institute of Directors, argues that this way of thinking starts far too late. The handbook, Crisis Communication for Boards: A Handbook for Good Governance, argues that crisis communication belongs within governance, risk management, and organizational resilience. By the time a crisis breaks, many factors that determine how well an organization responds have already been decided. Has the board understood the risks? Has it treated reputation as a strategic risk? Are escalation routes in place? Can uncomfortable information move quickly from the front line to senior management and the board? Have stakeholder relationships been built before they’re urgently needed? Has anyone rehearsed making decisions quickly under scrutiny with incomplete or conflicting information? The handbook’s point is that crisis management begins long before a crisis hits, and boards should own it as part of their responsibility for organizational resilience. Communication runs through crisis management because it shapes how events are understood, how decisions are interpreted, and whether trust is maintained or lost. One line early in the handbook captures the argument: “You cannot communicate your way out of a crisis that you behaved your way into.” That shifts the communicator’s role. If communications is brought in only after management has made its decisions, when the organization is already under pressure, the communicator starts at a disadvantage. Senior communicators should be involved earlier to help boards understand how decisions may be perceived, where stakeholder concerns are emerging, which assumptions may be wrong, and where the organization’s behavior may be creating reputational risk. That means treating communications as an intelligence function as well as a messaging function. The handbook emphasizes horizon scanning, stakeholder relationships, and the flow of information throughout the organization. Emerging risks increasingly cross legal, regulatory, HR, operations, technology, and communications boundaries. Communicators can help boards understand what’s happening and what it may mean for trust, legitimacy, and reputation. The handbook is blunt about the consequences of getting this wrong. Poor or delayed communication can itself become a reputational risk. Minor incidents can escalate, information can become trapped in silos, and boards can lose touch with what employees, customers, and other stakeholders are saying. Once trust is damaged, recovery can take years. It also addresses AI, cyber risk, misinformation, and the speed at which reputational issues spread. With AI, it asks familiar governance questions: Who decided? Who verified? What accountability was in place? Technology can change the nature and speed of a crisis, but it doesn’t remove the board’s responsibility for decisions. The handbook goes further, arguing that strategic communication belongs inside the governance structure, with access to the board and potentially representation at board level. A board that waits until a crisis to engage communications has already lost some of the ground on which trust is built. So, Shel, have we been using the term “crisis communication” too narrowly? Is it about what you say once a crisis happens, or about how the organization is governed before it happens? Shel Holtz: I think the answer is yes. Communication has to happen during a crisis, but it can’t be the work of a department brought in to clean up afterward. It has to be woven into the organization’s approach to crisis governance. That isn’t solely a communications function. The question of having a “seat at the table”—a term you know I hate—matters if an organization doesn’t take crisis preparedness seriously. Some leaders think, “We’re fine. We don’t have crises. If one happens, we’ll cross that bridge when we come to it.” If you can’t get access to leadership to make the case for crisis governance, then that lack of access is a problem. This is timely for me because I’m helping develop a crisis management plan. We’re calling it that rather than a crisis communication plan. I’m not suggesting you can necessarily manage every aspect of a crisis, but the plan needs a name. We’re developing it by committee, and in this case, that’s a strength. Crisis governance requires active participation from several parts of the organization. Communications can’t wave a document when a crisis hits and announce, “Here’s the plan, everybody.” People need to have helped develop it, understood it, and practiced it. Our committee includes a senior operations leader, our general counsel, our vice president of best practices, and the head of safety. They all have a say in how we approach crisis governance. Horizon scanning falls largely to me in communications, although leadership listens when something comes to us through our monitoring service or another channel. The plan also identifies who should convene on the crisis team based on the nature of the event. You need different expertise for an operational crisis than for a digital security crisis. Those teams need practice to develop the muscle memory to respond well. As our friend Philippe Borremans would say, we’re in an era of polycrisis. You may face several dimensions of one crisis at once, or two unrelated crises that begin to intersect. Adversaries can see what’s happening and take advantage of it. I found an example in our industry that started as a safety issue, became a legal issue, then a financial one, and ultimately put the company out of business. The ability to pivot quickly has to be built into both governance and training. That’s why tabletop exercises matter. Give the team a realistic scenario, put people through their paces, and introduce complications that force them to adapt. Ideally, you would do that four times a year. I don’t know how you build muscle memory if it’s merely an annual exercise people feel obliged to complete. None of this sits solely within a communications department’s purview. Communications informs the communication dimensions, but leadership and the other functions have to prepare together. It cannot begin with a call to communications saying, “We’ve got a crisis. Help.” Neville Hobson: Another point that stood out to me is that communicators should challenge decisions, not merely explain them. The handbook presents senior communicators as sources of stakeholder intelligence and constructive challenge. They can help boards anticipate how decisions will be interpreted, spot emerging reputational risks, and counter groupthink. That goes well beyond creating messages. If the communicator enters the room only after a decision has been made, have we misunderstood the role? Sometimes the most valuable contribution is to say, “You may not want to do this in the first place.” Shel Holtz: That connects to identifying issues before they become crises. I can’t remember the exact wording, but I recall a thought from the IABC Excellence Study: A crisis is what an organization faces when it fails at issues identification and management. If you see an issue coming and act, you may keep it from becoming a full-blown crisis. One definition I use for a crisis is that it can become existential. If it spins out of control, it can affect your ability to do business. The organizations that say, “We’ve never had one; things are going great,” may be particularly unprepared when a crisis arrives. Our plan includes holding statements—not just one, but several for different circumstances—that can be revised and issued quickly. Delay can look evasive. Victims, adversaries, and others may speak first, and that first account can become the version people remember. Then you have to overcome an impression that has already taken hold. Even if all you can say is, “We’ve just learned of this and are looking into it,” you can tell people when to expect an update. If you don’t have more information at that point, say you’re continuing to investigate and give them another update time. Give people a reason to turn to you for reliable information. Neville Hobson: The handbook is worth reading rather than skimming. We’ll link to the PDF on the CIPR website in the show notes. It includes case studies and a toolkit that may be useful in your planning. And I’ll repeat that line: “You cannot communicate your way out of a crisis that you behaved your way into.” Shel Holtz: Self-inflicted crises are among the hardest to handle. In some crises, the organization is a victim too, even if others are harmed along the way. One last thought: Many crisis plans fail to include employees among the stakeholders they need to communicate with. Employees can be your strongest advocates and supporters during a crisis. If you equip them with accurate information, they may be able to answer questions from clients, vendors, partners, consultants, and others they work with. Dan York: Greetings, Shel and Neville, and FIR listeners around the world. It’s Dan York on a glorious autumn day in northwest Vermont. I’m drinking warm apple cider and will soon head to an orchard for apple cider doughnuts. I want to talk about three things happening online. First is an effort announced this past week to close the language gap in AI systems. On September 21, the Gates Foundation announced a five-year commitment with about 60 organizations to make AI services available in more languages. More than 7,000 languages are spoken worldwide, but most online content is in perhaps seven to 10 languages. AI training data and interfaces are concentrated in even fewer. As part of its Goalkeepers report, the Gates Foundation looked at what’s needed for more people to benefit from AI, and language is a major piece of that. My employer, the Internet Society, signed on, as did the Coalition on Digital Impact, or CODI, whose board I sit on. CODI is a nonprofit focused on helping people navigate the internet in their own languages. There’s nothing concrete to report yet, but I think this is an exciting development. I’ll share more in the months ahead, including ways people might get involved. Everyone who chooses to use AI should be able to do so in their own language. My second topic is video. Bluesky now allows uploads of videos up to 10 minutes long. We’ll see whether that encourages more people to make videos there. YouTube also held its Made on YouTube event on September 23 and announced a range of updates. They include ways to customize the feed on your home page, an Ask YouTube tool for recommendations, more editing tools for Shorts, and features for creating vertical microdramas. YouTube is also adding live-streaming capabilities that appear aimed at competing with Twitch, including new ways for hosts to interact. It says live automatic dubbing into multiple languages is coming by 2027. There’s a lot for creators and communicators who work with video to explore. I’ve also been following smart glasses. I’m intrigued by their possibilities but concerned about privacy. Cameras built into glasses can be used to record people inappropriately, including women who have not consented. At the same time, I’ve seen how helpful the technology could be for people with disabilities. Someone who is blind, for example, might use smart glasses to navigate a store and learn what’s on the shelves. That promise comes with difficult questions. You may not know whether the person wearing Ray-Ban smart glasses nearby is livestreaming. Meta has announced more glasses, including a future VR model that sounds closer to goggles than ordinary eyewear. Snap has also announced new augmented-reality Spectacles, with a price around $2,200. Meta has introduced an audio-only option, which may address some concerns about cameras. I’m interested in what these devices can do, but I remain concerned about the privacy implications. Finally, Threads has announced podcast tools intended to help creators share shows, bring listeners back, and start conversations. They include profile banners, episode link cards, transcript and guest cards, new-episode reminders, and podcast insights. I have a Threads account but don’t use it much anymore. Still, if you’re a podcaster who uses Threads, these tools may be worth a look. I’ll send it back to you. I need to buy some apple cider doughnuts. Bye for now. Shel Holtz: Thanks, Dan. I’ve also been following the smart-glasses announcements. Meta’s glasses were around for a while before they developed a stigma, but that stigma has become real. You may remember Google Glass and people being called “Glassholes” for wearing it in public. Something similar seems to be happening with camera-equipped Meta glasses. A version without cameras strikes me as a smart move. I can see the value of glasses that display turn-by-turn directions or information about what I’m looking at. When I travel internationally and encounter a sign in another language, I pull out my phone and use Google Lens to translate it. Having that translation appear in my glasses would be wonderful. There are many useful things AI glasses could do without intruding on other people. As for recording, people already use devices such as Plaud and Fieldy. If you ask permission or let people know you’re recording, I don’t see a major issue. I hope AI glasses develop further, though I’m not inclined to buy a pair from Meta. I’ve also been reading about Meta’s new VR glasses. Initially, I was enthusiastic. I use a Meta Quest 3 mainly to work out with an app called Supernatural. There’s a wonderful story behind Supernatural that we should discuss on the podcast one day. I wanted to know whether my Quest library would work with the new glasses, and apparently it will. I also wondered whether glasses would stay on while I’m jumping around during a workout, and the answer seems to be yes. But then I saw the $1,200 price. People complained that the Quest 3 was expensive at $500. I won’t be rushing out to buy the new device. I’ll see what people say after it’s been out for a while; my Quest 3 works just fine. Great report, Dan. Thank you. Dr. Eleni Nicola is described as an art therapist from Cyprus. She has been quoted more than 30 times in recent months: in Forbes on resetting your home office, Vice on what color to paint your walls to calm an anxious dog, Glamour on crying therapy, and Parade on the right way to retire. Different outlets have called her a clinical psychologist, an art therapist, or a creative wellness expert. The problem is that she doesn’t appear to exist. Press Gazette reported that an image detector rated her profile photo 10 out of 10 for being artificial and her quotes as 100% AI-generated. Her LinkedIn profile was created late last year despite claims of a long career. She isn’t listed by the Art Therapy Credentials Board, which lists thousands of art therapists. Every one of her online profiles links to the same paint-by-numbers kit retailer, DeVincify, where she is presented as an “expert in residence.” This has become a business model. A quote in Forbes brings a link from a high-authority news site. That can improve search rankings, and rankings can drive sales. SEO operators are building fake personas as part of link-building campaigns. When Press Gazette asked DeVincify to verify her, the company’s statement about taking the matter seriously was itself identified as AI-generated. At the time of publication, her profile was still active. This is far bigger than one case. Press Gazette has identified more than a thousand articles across outlets including Mail Online, Metro, The Sun, The Mirror, HuffPost UK, Yahoo, and The Independent that relied partly or entirely on fake experts. More than 600 were traced to a single company. One fictitious psychologist was quoted hundreds of times before anyone established that she wasn’t real. The fakes aren’t limited to sources. Tech sites have pulled articles attributed to finance writers whom Press Gazette couldn’t verify. One claimed a Deloitte career that Deloitte has no record of; another used a headshot that appears to belong to a computer repair technician in the United States. Dan Simon, who runs the journalist-sourcing platform Qwoted, says the platform handles more than 15 violations a day and blocks about 50 prospective clients from registering each day. As he put it, “AI has just poured petrol on the disinformation fire.” The combinations include real agencies with fake experts, real experts with fake expertise, and fake agencies with fake experts. David Higgerson at Reach, publisher of The Mirror, said agencies Reach regularly works with had supplied fake experts without realizing it. They had sourced the quotes in good faith. That makes PR both a victim and a vector. If an agency fails to verify the experts it puts forward, it can launder fakes into the news. These are things we need to catch before publication. A LinkedIn profile is no longer enough to establish that someone is real. Reach has whitelisted hundreds of PR firms, built an AI tool called The Detective to assess whether spokespeople are genuine, and added another layer of verification for people quoted on health or money. Qwoted is introducing video verification. One platform requires responses to be recorded live on camera. Another stopped offering open access because verifying everyone would otherwise have threatened its business. Simple checks still help. Ask for a five-minute call. Ask the person to respond to a specific photo request. If they claim a professional credential, check the relevant registry. One supposed expert stopped responding when asked for a photo. Another was exposed when an agency checked a WhatsApp profile picture and found that it belonged to a man doing SEO for a recipe blog. Neville, what else can an honest communicator do to avoid perpetuating quotes from fake experts? Neville Hobson: I was surprised by how readily fake experts were accepted by major media outlets. I remember a case involving The Telegraph a few months ago. I can’t recall the details, but it published something false, then removed it and apologized. Why wasn’t it checked first? We make the same argument about people using AI-generated content without checking it, but source verification is even more fundamental. I don’t know precisely how every newsroom chooses and verifies the people it quotes. Clearly, though, something is seriously amiss if this is happening on an industrial scale. We may need stronger ways to prove digital identity, comparable in some respects to the checks involved in applying for a passport or, in my case as a school governor, undergoing a criminal-records check. Informal methods aren’t working, and manual checks are difficult to scale. The people creating fakes can scale their efforts. We may have to give up some of the casualness with which we’ve approached verification. Shel Holtz: One detail that struck me was that this involved a paint-by-numbers company. If a company selling paint kits is doing this, who else is? We tend to picture shady crypto operators or other familiar bad actors. But anyone trying to sell something and gain an SEO advantage might be tempted. A journalist’s credibility is at risk when a quoted expert turns out not to exist. News organizations need a process for vetting sources, and they should tell readers what that process is. PR people need one too. If you use a fake expert’s quote to support a press release or offer that person to a reporter, why would that reporter trust you again after discovering the deception? This reminds me of a story we discussed years ago about a purported bank robbery in Texas. A local publication reported it, a larger outlet picked it up, wire services followed, and it spread through news sites, television, and radio. The original story was fake. Each outlet assumed the previous one had checked it. Now the assumption is, “This person has ‘Dr.’ before their name and a LinkedIn profile, so they must be legitimate.” That is woefully inadequate. We need processes to catch these fabrications before we send them into the world. Neville Hobson: That’s the issue. It doesn’t make much sense for every media outlet to build a different verification process. That won’t scale. There may be a role for independent services whose job is to verify that people are who they claim to be. Imagine a PR firm with a strong reputation for providing knowledgeable spokespeople. If just one turns out to be fake, doubts could arise about everyone the firm has supplied. Firms need a way to bolster their own credibility and give journalists and clients confidence in the people they put forward. AI will only make this problem worse. It’s unfortunate, but we’ve reached a point where verification is essential. Shel Holtz: I like the idea of vendors offering verification as a service. It’s also an example of new work arising as AI changes existing jobs. In the meantime, a five-minute video call or a specific photo request can help establish that an individual is real. Those checks may not scale across thousands of sources, but they are steps people can take while better systems are developed. Ultimately, we need scalable solutions. This affects reputation and trust, particularly if people act on advice that was fabricated simply to sell a paint-by-numbers kit. Neville Hobson: Amazing. Shel Holtz: That’s our report for today. We would love to hear your comments. Audio comments are especially welcome—you heard one today. Send us an MP3 of up to three minutes, or record a message directly on the FIR website by clicking the “Send Voicemail” tab on the right side of the page. You can also comment where we post episode announcements, including LinkedIn, Facebook, Threads, Bluesky, and Mastodon, or leave a comment in the show notes at FIRPodcastNetwork.com. We’ll find it, and you can be part of the conversation. Ratings and reviews mean a great deal to us, too. Wherever you get your podcasts, we’d be grateful if you would rate and review the show. We plan to record our next monthly episode on Saturday, October 24. Neville Hobson: That’s right, and we’ll publish it on Monday. Shel Holtz: Monday, October 26. As for shorter midweek episodes, we’ll see how Neville is feeling. If you subscribe to the show, new episodes will appear in your feed as soon as they’re available, without your having to wait for an announcement. And that’s a wrap for this episode of For Immediate Release. The post FIR #529: Fake Quotes from Fake Experts Find Their Way Into Top-Tier Media appeared first on FIR Podcast Network.

49 MIN

SEP 3, 2026

FIR #528: Why Marketers Are Suddenly Obsessed with Fandom

Fandoms are not synonymous with audiences or communities. The difference: Fans’ identities are tied up with the team, artist, product, or other object of their passion. It’s not surprising that marketers are chasing fandoms, in hopes of turning their brands into something people are genuinely excited about. But not everything is a potential flashpoint for a fan base. In this short midweek FIR episode, Neville and Shel explore what distinguishes fans from members of other audiences or communities, how brands might leverage them, the wisdom of trying to spin up a fan base, and why fandoms might rebel against attempts to use them for marketing purposes. Links from this episode: WTF are fandoms, and why are marketers chasing them? Consumers as Brand Fans: A Systematic Literature Review and Research Agenda Using the TCCM Framework Social media and fandom How community is redefining Media & Creative in 2025 The next monthly, long-form episode of FIR is tentatively scheduled to drop on Monday, September 28. We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected]. Special thanks to Jay Moonah for the opening and closing music. You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog. Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients. Raw Transcript: Neville Hobson: Hi, everyone, and welcome to For Immediate Release. This is Episode 528. I’m Neville Hobson. Shel Holtz: And I’m Shel Holtz. There is a new word making the rounds in marketing circles these days, and that word is “fandom.” Obviously, fandom isn’t anything new. Beatles fans constitute a fandom. Star Trek fans are a big fandom. Sports teams have had fandoms for as long as sports have been around. What’s new is the idea that fandom should be a marketing strategy. Marketers are paying attention because a fandom represents something considerably more valuable than an audience. An audience is passively watching you. That notion led Jay Rosen to write a 2006 post titled “The People Formerly Known as the Audience,” describing the shift away from a one-way mass-media model to a networked model in which people who once passively consumed information can now publish, respond, organize, remix, and distribute it themselves. That’s not always necessarily a good thing, but it is what it is. A community gathers around a shared interest, but a fandom goes even further. Fans incorporate that interest into their identities. They create things, develop rituals and inside jokes, recruit other people, recommend things to one another, and sometimes defend the object of their fandom with remarkable enthusiasm. One definition I particularly like comes from an article in Digiday: Fandom is where attention turns into identity. You can probably understand why marketers find that appealing. Advertising can buy attention. It’s much harder to buy identity. It’s harder to buy belonging, advocacy, or voluntary engagement. There is some evidence that fandom deserves to be treated as something distinctive. A systematic review published last year examined 38 marketing studies involving brand fans and concluded that fandom has moved from the margins of consumer culture into the mainstream. Brand fans really do behave differently from people who merely buy a product regularly. There is another reason fandom has suddenly become attractive: The media environment has fragmented spectacularly. People aren’t gathered around the same handful of television networks, newspapers, magazines, and radio stations anymore. They’re scattered among TikTok creators, YouTube channels, Reddit communities, Discord servers, gaming platforms, streaming services, group chats, podcasts, and sports communities. We could keep listing niches all day long. Instead of asking, “How do we reach women ages 25 to 44?” marketers can potentially find a group of people who are already intensely interested in something relevant to the organization. That changes the communication challenge. You don’t approach a fandom as an audience waiting to receive your message. You’re entering somebody else’s existing culture. That culture already has leaders, vocabulary, expectations, and history. It has ideas about who belongs and who doesn’t. Increasingly, creators are among the people who have earned credibility within those cultures. The strategic question isn’t, “Which influencer can deliver the biggest reach?” It becomes, “Who actually has standing with these people, and do we have a legitimate reason to be here?” That last question is one I think organizations should spend much more time considering, because fandom is in danger of becoming a bandwagon. Whenever marketers discover something people genuinely care about, there’s a temptation to turn it into a technique. Community, authenticity, purpose, and engagement have all become marketing techniques. Once every agency presentation starts promising to “activate fandom,” that’s probably the time your Spidey sense should start tingling. You can’t manufacture a fandom just by announcing that you have one. Nor is every organization capable of inspiring fandom. Nobody necessarily wants to build part of their identity around an electric utility, payroll software, an insurance company, or the manufacturer of the toner cartridge in the office printer. How many true HP fans are there, really? That doesn’t mean those organizations can’t have excellent customer communities. They can create useful resources, cultivate advocates, involve customers in product development, and provide terrific experiences. There are plenty of organizations that do these things well. But calling every highly engaged customer a fan doesn’t make fandom a meaningful strategic concept. It just gives customer engagement a trendy new name. There is another danger: Brands may identify an existing fandom and assume its passion is available for rent when it probably isn’t. Fans can tell when a company has wandered into their community because somebody in marketing spotted a cultural trend. Borrowing the language, memes, or symbols of a fandom without understanding them can make a brand look less relevant, not more. The more intensely people identify with the community, the more protective they may be of it. After the animated television series Rick and Morty made McDonald’s discontinued Szechuan sauce into a fan obsession, McDonald’s embraced the phenomenon and announced a one-day return. But stores received extremely limited supplies. Fans lined up for hours, and some traveled considerable distances, only to find that no sauce was available. The resulting anger produced protests and the hashtag #GiveUsTheSauce, and McDonald’s ultimately had to apologize. The mistake wasn’t offensive cultural appropriation. It was a company correctly spotting a fandom but badly misunderstanding its intensity—and failing to deliver the experience it had encouraged. I like one test suggested by the material I was reading: If the brand disappeared from the activation, would fans still say that what happened made their community better? That’s a pretty demanding standard. Your contribution might be access, useful information, an experience people couldn’t otherwise have, a genuine role for fans in developing something, or help that enables creators to do better work. Merely inserting your logo probably doesn’t qualify. It also means fandom isn’t likely to be an especially good short-term campaign strategy. These relationships require cultural fluency and sustained participation. If the budget disappears after the quarter ends or management expects an immediate conversion metric, conventional marketing may be the more sensible choice. There are plenty of examples of companies getting this right. LEGO supports its AFOLs—Adult Fans of LEGO—and an existing culture rather than pretending it invented that culture. Harley-Davidson didn’t merely accumulate loyal customers. It helped riders organize around a shared identity by creating the Harley Owners Group, or H.O.G. It provides chapters, rallies, riding challenges, a magazine, events, and other opportunities for motorcycle owners to interact with one another. Even the pumpkin spice latte gave Starbucks an opportunity to turn a seasonal beverage into a fall ritual. The company created a private Leaf Rakers Society Facebook group for enthusiasts, which grew to more than 26,000 members. Starbucks didn’t invent the pumpkin spice latte ritual or the anticipation surrounding it, but it certainly learned how to play along. What should communicators do with all of this? First, don’t begin by asking, “How can we build a fandom?” Start by listening. Find out whether communities already exist around your organization, its products, the problems it solves, or interests adjacent to it. Second, distinguish an audience from a community and a community from a fandom. Don’t inflate ordinary engagement numbers into evidence that people have incorporated your brand into their identities. Third, if you enter an existing fandom, learn its culture before trying to influence it. Identify the people who have earned trust there, including creators, and involve them rather than treating them as distribution channels. Fourth, look for ways to enable participation rather than merely generate impressions. Give people opportunities to contribute, create, influence, and connect with one another. Finally, measure whether the relationship persists when the campaign stops. Are people coming back? Are they creating things voluntarily? Are they recommending you? Are they talking to one another rather than just talking to you? Ultimately, fandom isn’t something a communication department creates. It’s something people decide they belong to. If organizations remember that distinction, fandom could be a genuinely useful way to think about relationships. If they don’t, it may become the next marketing buzzword we spend a couple of years enthusiastically putting into PowerPoint decks before moving on to the next one. Neville Hobson: You addressed this a bit, but what’s the difference between a community and a fandom? When does one become the other? Shel Holtz: The key differentiator is that the members of a fandom incorporate the object of their interest into their identities. It’s no secret to anyone who knows me—and probably half the listeners to this show—that I’m a Deadhead. I love the Grateful Dead. Grateful Dead merchandising makes a ton of money, and I buy some of that stuff. I recently received a scale model of the Wall of Sound, the massive audio system the band toured with in 1974. It’s 3D-printed and even has a little Wi-Fi speaker built into it. It was a birthday present, and I just love looking at it. It’s part of that fandom. I’ve read an entire book about the Wall of Sound. I’ve probably read 20 books about the Dead. I was listening to a 1985 concert on my drive to the office this morning. I am a fan. That is part of my identity. The Grateful Dead isn’t just a band I like and whose music I sometimes listen to. I may be part of a community of people who like Elvis Costello, for example, but that doesn’t make me part of an Elvis Costello fandom. I don’t incorporate that particular artist into my identity. That’s the real differentiator. Neville Hobson: That’s a good explanation. Identity is the key, based on what I’ve been reading as well. Are marketers suddenly so interested in fandom because traditional ways of reaching audiences are becoming less effective? Shel Holtz: That’s my take. Marketers are desperate because some traditional means of reaching and engaging people have vanished, some have transformed, and most have fragmented. They’re looking for magic bullets. You read one or two articles about how an organization leveraged a fandom, and suddenly the response is, “We’ll turn our customers into a fandom, too.” But I just can’t see someone becoming a fan of their electric utility. You’re a customer. You may even be an enthusiastic customer for some reason, but I’m never going to incorporate the PG&E brand—Pacific Gas and Electric—into my identity. Ever. Neville Hobson: Here’s a slightly rhetorical question prompted by your comments about Deadhead fandom: Can a brand genuinely become part of a fandom, or does the relationship become transactional the moment the brand begins mapping and monetizing it? Shel Holtz: I think both things can be true. A brand can become part of a fandom if it approaches the opportunity in the right way. First, there has to be a genuine connection to that fandom. I’m trying to think of something that would connect naturally to the Grateful Dead fandom. It might be a guitar brand or one of the nonprofits affiliated with the band, such as Rock the Vote. If you listen to the Good Ol’ Grateful Deadcast, the band’s podcast, it has done a deal with Dogfish Head Brewery on the East Coast. The brewery makes Grateful Dead-branded beer. They have thought it through and developed the connection and relationship so they can reach the fandom through the podcast. That may lead people to flock to the beer and talk about it. It doesn’t necessarily mean they’ll become fans of the beer. They may become consumers of it, perhaps even regular consumers, but the company is leveraging an existing fandom. I think there are also opportunities for brands to build fandoms. Some happen naturally. Consider a car brand. There are Mustang fans. They love their Mustangs. They attend Mustang rallies and car shows. That reminds me of the brouhaha we discussed many years ago, when Ford sent a cease-and-desist letter to a fan who ran a discussion board. It wasn’t about the Mustang, though, was it? It involved a pickup truck. Neville Hobson: I think it was the pickup—the Ranger. Shel Holtz: The Ford Ranger, that’s right. The Ranger Station was the name of the site. Ford sent the cease-and-desist letter, and the site’s owner posted a note saying, “I have to shut this down. I received a cease-and-desist letter from Ford.” The fandom erupted. This was an organization that had a fandom but didn’t recognize its power. Ford could have been leveraging it to expand the customer base. There are opportunities to identify existing fandoms that may revolve around your products, services, or company. More often, though, it will be something adjacent to your business that you can tap into. You need to understand the culture. You need to know who drives it and understand its rituals and jargon. You can’t parachute in and announce, “Here we are. We’re going to tap into this fandom. Aren’t you excited that we’re here?” People will simply say, “No, we’re not. Go away.” It’s definitely worth considering, but it isn’t a magic bullet. For most organizations, I think it’s ridiculous to assume they can develop a fandom around what they do. Neville Hobson: I agree. We’re touching the edges of a discussion about a shift from authenticity—which is unquestionably overused in marketing—toward reciprocity. Does a brand have to earn the right to participate in a fandom? If so, what does it have to give back? There’s an interesting communication principle underneath that question: Don’t begin by asking what value the community can create for the organization. Ask what value the organization can create for the community. What do you say to that? Shel Holtz: Absolutely. That’s how you build credibility and relevance within a fandom, unless it’s a built-in fandom involving something like the Ford Ranger or Ford Mustang. If you have identified an existing fandom that isn’t centered on one of your products but has a genuine connection or relevance to your business, it may be worth investing the time, effort, and money required to ingratiate yourself with that community. How do you do that? You might make access to the object of the fandom a little easier or give fans access to the right merchandise. Ask what you can do to make their fandom more meaningful or more fun. Then you may be accepted, and fans may consider doing something for you—which is obviously one reason you became involved in the first place. You have to be strategic rather than simply jumping in and saying, “We connected the dots. Look at our connection to this fandom. Don’t you want to help us sell our product?” I fear that’s what will happen with a lot of marketers who read about this and think, “There’s a solution for me. I’m going to jump all over fandoms.” Neville Hobson: I struggle a bit to see the precise gap between community and fandom. When does one become the other? I understand the examples you’ve given: the Grateful Dead fandom and The Ranger Station, for instance. Shel Holtz: And the Adult Fans of LEGO. Neville Hobson: That one, too. But do these groups actually refer to themselves as fandoms? What do their members call themselves? Shel Holtz: Well, they call themselves Adult Fans of LEGO. The “F” stands for “fans,” so they do. Would members of Harley riders’ groups—the H.O.G.s—call themselves a fandom? Probably not. They would probably call themselves H.O.G.s, but they’re definitely a fandom. I remember working for a company in the early 1990s whose vice president of compensation and benefits was a Harley owner. He rode with his club every weekend and wore a Harley jacket to the office—and this was a pharmaceutical company in the ’90s. He was a fan. There’s no question about it. He wasn’t merely a member of a Harley community. Being a Harley rider was part of his identity. Neville Hobson: This is rich territory for marketers to get wrong. Forgive me for beginning with the cynical perspective, because I can also see the opportunities available to organizations that do it well. But are we trying to manufacture something that doesn’t need to be there? I don’t particularly like the word “fandom,” although I’m not part of a fandom myself. I belong to many communities, but I’m now wondering about the difference for me personally. Do I see any of those groups as part of my identity? Not really. I probably haven’t encountered one that produces that “wow” for me. Is fandom a manufactured opportunity for marketers, or is it genuinely something worth investing in? Shel Holtz: Again, I think both things can be true. We’re going to see a lot of marketers try to manufacture fandom because it’s a buzzword making the rounds. But some organizations will have a lightbulb moment and realize, “We have a fandom. We’re just not leveraging it.” Imagine if Harley-Davidson had never organized local chapters or created clothing people could wear to express their fandom. What if the company had said, “We sell motorcycles. If our customers want to form clubs and ride around, that’s fine—but they had better not make clothing with our logo because that’s a copyright violation”? Harley-Davidson probably wouldn’t have the identity it has today if it hadn’t recognized that it had a fan base it could support and leverage. The same is true of Adult Fans of LEGO. The company produces kits that clearly aren’t intended for little kids snapping together a few bricks. Some contain thousands of pieces and cost hundreds of dollars. Those are definitely products for adult fans. As long as you listen to fans and nurture the relationship, there can be value in it. Fans have a relationship with the object of their passion. Organizations need to examine that relationship and ask: Are we strengthening it? Are we reinforcing it? Or are we merely trying to exploit it? Neville Hobson: Exactly. Shel Holtz: That’ll be a 30 for this episode of For Immediate Release. The post FIR #528: Why Marketers Are Suddenly Obsessed with Fandom appeared first on FIR Podcast Network.

23 MIN

AUG 24, 2026

FIR #527: Delegating Hiring Decisions to AI

94 MIN

AUG 17, 2026

FIR #526: Forget Anthropic’s AI Watermark. Can You Defend Every Sentence?

Anthropic’s watermark scheme is the focus of so much discussion that you could be excused for thinking there was nothing else to talk about. For many, it’s the solution to identifying all those evil-doers who offload their writing to large language models. But we are wasting far too much time trying to determine whether (and to what degree) AI was involved in creating content. Much more important is determining whether the content was crafted with respect for the reader, and whether the creator can stand by every word. That’s the idea behind the new AI writing policy from the software company Clay, which makes far more sense than trying to detect a watermark. Links from this episode: Clay AI Writing Policy: 4 Guiding Principles How Claude marks AI-generated content Anthropic’s watermark survives copy-paste, but not the real dev workflow Can Anthropic’s invisible watermarks curb ‘AI slop’? Researchers remain sceptical Anthropic’s text watermarks signal new front in AI detection Claude’s new Scarlet Letter watermark is invisible—for now How to Build a Responsible AI Writing Policy How Claude’s text watermarking works Claude is now watermarking every response — here’s what that means if you use AI for writing I’m Begging You: Never Write With A.I. The next monthly, long-form episode of FIR is tentatively scheduled to drop on Monday, August 24. We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected]. Special thanks to Jay Moonah for the opening and closing music. You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog. Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients. Raw Transcript: Neville Hobson: Hi, everyone, and welcome to For Immediate Release. This is episode 526. I’m Neville Hobson. Shel Holtz: And I’m Shel Holtz. In the last episode of FIR, we talked about workslop and the implications workslop brings to the workplace. We’re going to continue down that path today. I’m sure you’ve all seen the headlines about Anthropic putting an invisible watermark on anything Claude writes. I want to separate what that actually does from what people have been claiming. First, you won’t see this watermark. It’s not a “written by Claude” tag. It’s machine-readable. It survives copy and paste, and it may even survive some editing. The reason Anthropic came up with this is to comply with the transparency rules under the EU AI Act, and it applies everywhere Claude runs: the app, the API, Claude Code, Cowork, cloud providers — you name it. You can’t paste text into a public detector and expose someone yet. Anthropic says detection tools are coming, but it hasn’t released any of them. Here’s roughly how it works: Every time Claude generates text, it’s choosing between several equally good next words — say, “gray” or “overcast.” Normally, that choice is random. With watermarking, though, a secret cryptographic key nudges that randomness in a consistent way. No single word looks suspicious, but across a couple hundred word choices in an article, the pattern becomes statistically detectable if you have the cryptographic key. That’s completely different from tools like Pangram, which look for writing patterns that seem AI-like — you know, em dashes, overuse of words like “delve” or “tapestry” or “underscore,” constantly grouping things in threes. These are all things real writers do, by the way. The rule of three is nothing unique to AI, nor are em dashes. That’s why I don’t put much stock in these tools. Anthropic’s detector doesn’t guess. It tests whether the text matches the pattern its own cryptographic key would produce. Now, it’s particularly important to understand that a detected watermark means the content was processed by Claude. It doesn’t mean Claude wrote it. You could write something yourself, have Claude clean up the grammar, and it would still carry that watermark. Axios flagged exactly this risk for communications teams that polish a human-written press release with Claude. It also works in reverse, by the way. Heavily edit, paraphrase, translate, or blend the text with other writing, and that watermark can disappear. Short passages may not carry enough signal to be detected at all. So this isn’t a foolproof “Did a human use AI?” detector. Picture a reporter running a company statement through a detector and calling it AI-generated when your team actually wrote it and just had Claude do the final polish. The watermark tells you about processing. It can’t tell you who did the thinking. And that brings me to Clay, the software company, which just rolled out an AI writing policy. I think this is actually more important than a watermarking system. The policy had its origins in Clay’s engineering department, but it has since gone company-wide. It’s not a ban. Brainstorming with AI, drafting, proofreading — all that’s fine. Instead, the policy has four principles. First, stand behind every idea and every sentence. Second, writing is thinking. Third, spend more time creating a document than you expect someone to spend reading it. And fourth, longer isn’t better. And AI, by the way, notoriously pads its content. I like this policy a lot. I like it a lot more than “Don’t use AI to write.” Because sometimes AI writing is exactly the right call. Plenty of people are great at their jobs and bad at writing. Engineers are a great example. I’ve used engineers as an example of this before. If AI helps these people organize their explanation and turn an incomprehensible email into something readable, that’s good for everyone. The problem is AI can make bad thinking look like good writing. This is what we were talking about last week with workslop. You give the model a thin prompt, it hands back three polished pages, and now you’ve just shifted the work of figuring out what you meant onto everyone else who has to read it. The prose looks finished, so the creator of that content feels like they’ve finished, but the thinking never actually happened. That was how we defined workslop last week when we talked about it. So companies don’t need an AI policy that’s focused on AI. They need one that’s focused on accountability, quality, and respect for the reader. Cover the basics, for sure. List the approved tools. Talk about confidential information and how it gets used. Talk about fact-checking and human review and brand voice and blah, blah, blah. But add that Clay test: Can you defend every sentence? Does it represent what you actually think? Did you verify the facts? Did you cut the padding? And did AI make the communication better, or did you just shift your effort onto your audience? For communicators, that last question is the one to focus on. The watermark debate is going to tempt organizations to obsess over detection: Was AI used? Can we prove it? Should we disclose it? Yeah, sometimes, I guess, that’s a fair question. But the better ones are: Is the thinking ours? Is it accurate? Does it serve the reader? And is a human willing to stand behind every word? If we can get those things right, I’m not too worried about whether Claude helped fix a few sentences along the way, watermark or no watermark. Neville Hobson: Hmm. We have talked about this before — this kind of weird obsession so many people seem to have with trying to figure out whether someone used AI to write a piece so they can exclaim with great glee, “Yeah, this guy wrote this. It’s 96 percent AI. He didn’t write it at all. It’s a scam, it’s fake,” blah, blah, blah. Shel Holtz: Yeah. Neville Hobson: We’re already seeing it happen with this. Someone wrote the other day about this, “Finally, a way to get rid of AI slop.” What? I mean, isn’t — Shel Holtz: No. Neville Hobson: It isn’t going to do that. This obsession isn’t going to stop, I don’t think. And, in fact, maybe the best thing out of the Clay principles is taking the focus away from that and moving it to the writing, to authorship, to accountability, as you point out. The caution, I would say, is don’t obsess about this. Don’t set concrete rules that are so inflexible that you’re going to end up with writing that isn’t flexible at all. The point, I guess, is to concentrate on the actual writing. Think about what it is that you’re writing. So let’s move away from “Did you use AI to write this?” to something like, “What was your contribution to this?” That, to me, is a healthier discussion to have. There are still going to be lots of people who do the “gotcha” talk. I saw one today on LinkedIn: “I would never hire someone who uses AI to copywrite. That kind of approach isn’t for my team,” and stuff like that. I think asking what your contribution was encourages accountability, whereas “Did you use AI to write this?” is all about concealment and policing — discovering that someone did and assuming or implying that they’ve cheated somehow. That said, I think the point that you need to stand behind every idea and sentence is absolutely spot on. And that’s not new. We should have been doing that all along. That’s probably the right way to approach it. For example, don’t ask Claude to write something: “Here’s a topic. Write a 1,200-word piece arguing that communicators need to take responsibility for AI governance.” Claude produces it, you might edit two or three sentences, and you’re done. You publish it under your name. That is substantially AI-generated writing, and the watermark is relevant. But for what, though? So someone could say, “Gotcha”? You don’t know what that was for. You don’t know what’s wrong with that. Has someone deceived you? I suppose, by implication, if it’s under their name, they have deceived you in that they wrote it. They didn’t; the AI wrote it. But what’s the difference between that and, as a comparison, Grammarly or something like that, which suggests paragraph changes and you accept every single one of its recommendations? You end up with something that’s then 70 to 80 percent Grammarly, as opposed to you. Better, though, to avoid ethical issues and accusations and all that stuff, is that you do the writing. Don’t dump the stuff on the AI. You do the writing and introduce your own thinking to this. Have the AI — Claude or whatever it might be — check it. Ask it to proofread it. Ask it, “Is there any other angle I could have incorporated in this?” or “Do you think this is the right approach?” I do that a lot with the stuff that I write. And I must admit, probably two out of three times, I’ll accept some of the recommendations the chatbot comes back with. Sorry, the AI assistant comes back with. I don’t call it chatbot. Shel Holtz: Ha ha ha. Neville Hobson: And there I just slipped up. I did. So I think we’ll have to weather the gotchas all the time, and so be it. But if you are confident in doing the writing, using your AI assistant as the guide for you — as, in a sense, the editor that sits by your side, the critiquer that tells you what it thinks and where you could improve this or don’t say that — follow guidelines such as Clay has done, and you should be okay. Shel Holtz: Yeah. The piece on Clay’s writing policy makes the point, as I mentioned, that writing is thinking. Writing is the way we process our thoughts and test our thoughts. And this is a point that I think a lot of people have made. In fact, I just read this in The New York Times. I think it was over the weekend. I think it was an op-ed that was exhorting people to please, for God’s sake, write your own stuff because we don’t want to lose the ability to think. The problem is, I think most of these proclamations come from writers. You and I are writers, and the communities that we interact with on LinkedIn in particular are probably also writers. We’re connected to people who are engaged in communications, and hence you get the opposition to using AI to write. But again, what about an accountant? What about an engineer? There are so many jobs out there where being a really good writer was never a requirement when these folks were earning their degrees or their certificates. And now, because they’re in the business world, they have to communicate effectively. So how do you go about the thinking process? No, you don’t want to delegate it to AI. You don’t want to say, “Write an email about X” or “Write a blog post about X.” You want to think it through. But does that mean you think it through by writing? Well, not if you’re not a writer, necessarily. You may think it through by creating an outline, by doing a brain dump. It could even be — this is something Chris Penn talks about a lot — just recording your thoughts as audio and then uploading that file and saying, “These are my thoughts. Now turn this into a coherent email to this audience designed to produce this result.” Then you go back and you review it and edit it so that you can defend every sentence, every word. Again, I think this is all about respect for the reader. And if the way you demonstrate respect for the reader, knowing that you’re a terrible writer, is using AI, that’s appropriate. If you just knock out this email, people are going to go, “What is he talking about? I don’t know what he’s trying to get at here.” Respect for the reader is using the AI to make that message clearer, more cogent, more understandable, and more actionable. So, again, I really like those four pillars. And I think there undoubtedly will be ways to defeat the Claude watermark. People will come up with them. Anything that is designed to catch people is something somebody else is going to come up with a way to circumvent. In the meantime, though, as you say, this is not a surefire way to catch somebody. It’s going to identify that Claude processed this, not that Claude wrote it. So I think people need to calm down and start thinking about this tool in a way that provides the high-quality content that you’re trying to deliver to people — not just something polished that makes them have to figure out what you intended because it’s polished but no thinking went into it. The thinking still has to go into it, whether Claude’s going to do the lion’s share of the writing or not. Neville Hobson: Yeah. I mean, I think the interesting distinction is between AI-generated and AI-assisted writing. Although I add my own caveat to that, which is: Who cares? I mean, truly, do I want to get into a discussion about that? No, I do not. So I know the difference. I’m not evangelizing that everyone should understand and follow the difference, although it’s helpful to. For example, using Claude, you give Claude a short prompt — it’s not Cowork, by the way; this is just the chat — and it produces an article. You tell it, “I want to write about X, and the topic is this, and I want to achieve this. These are the points I want to make.” Maybe not even as much as that. And Claude produces, you know, a 600-word draft, and you lightly edit it. Lightly meaning maybe you change not the syntax so much, but certain expressive words that you might use. If that’s your bag, sure. Although you would have given — Shel Holtz: Or adding the Oxford comma. Neville Hobson: — your AI assistant the guidance about your preference on the Oxford comma. Shel Holtz: Presumably, yeah. Neville Hobson: It would know. So you do that. On the other hand, if you’ve already developed the argument, you’ve already got a clear picture in your mind, you’ve researched it, you’ve used Claude to challenge or improve your thinking, and you make the editorial decisions yourself, that’s another matter entirely. That, I believe, is AI-assisted writing. So, as we mentioned earlier in the discussion, the better question isn’t, “Did you use AI?” but, “What was your intellectual contribution?” So, writing is thinking. Yeah, I don’t disagree with that. Although, again, I don’t want to get hung up on having definitions all over the place about this. The danger is real of turning this into another policing technology, which many people are already starting to do. I did see in Anthropic’s FAQ about watermarking where they say, in answer to the question, “How do I check if a piece of text was written by Claude?”, “We will soon be offering a watermark detection API. We’re in the process of working out the details of its implementation.” So that’s coming. That’ll give those gotcha people a lot more ammo to say “gotcha” a lot more, I bet. It just takes attention away from what really matters with all of this. But, you know, like a lot of things that are new, we have to go through all this. So my recommendation is: Don’t give it too much attention. Be sure in your own mind that what you’re doing is, in your definition, the right way of going about it. You feel confident that it is. You have guidelines to follow, such as Clay’s, for instance, so you are able to say, “Yep, these four things — I do these things in all my approaches to this.” In which case, you can wave two fingers at the gotchas. Shel Holtz: Yeah, and I can’t emphasize this enough: The detection is not going to be focused on whether Claude wrote this. It’s whether it processed it. So you could have written it and then sent it through Claude for grammar, spelling, and punctuation, and the watermark will show up, even though you wrote it. So the gotcha is, I think, a little disingenuous in a lot of cases. I mean, it could be that somebody used it to write, for sure, but it’s not a sure thing. It’s not a lock that if the detector says, “Yes, Claude processed this,” then, “Aha! You wrote this with AI.” Not necessarily. It doesn’t mean that at all. Neville Hobson: But even if he did, so what? If someone says that, do I care? Not at all. Shel Holtz: Well, and again, I come back to the engineer or the accountant who wants to be understood and is not a writer. I think that’s where this tool shines and where we have opportunities for better clarity and better understanding in the workplace. Do I expect my professional writers to write? Yeah, absolutely. Do I expect my accountants to write? My lawyers? Not necessarily. Neville Hobson: Yeah. I mean, I think there’s one other thing about this, too. I did read Anthropic’s technical paper describing what this is and how it works. I’m going to have to actually ask Claude to simplify this — give it to me in simple terms so I can understand it well. Shel Holtz: Ask Claude. No, better yet, ask ChatGPT to do it. Neville Hobson: Because it talks about, for instance, let’s say you asked Claude to write a piece. You briefed Claude, it did that, you edited it, passed it back to Claude, it made some recommendations, which you implemented, including removing a chunk. You passed it back to Claude again. You went to and fro a bit, and you might then have — let’s say it’s an article for an academic journal, for instance — passed it to a colleague, saying, “Can you review this and give me your opinion?” They did, and you made some further changes as a result. The end result of all of that is so muddy that Claude would have — or rather, the watermarking wouldn’t be clear as to who wrote which bits and so forth. You could pinpoint where the words came from, but you couldn’t pinpoint who the author was. In which case, the gotcha folks are not going to be happy with that. But, you know, just get on with it, for God’s sake, and stop all this stuff. Shel Holtz: Yeah. And you don’t know which words it selected to accommodate this watermarking. If your edit changes them, then the watermark vanishes. So again, no guarantees here at all. This is compliance with the EU AI law. That’s all it is. By the way, I think you can expect to see the other frontier models follow suit, also to be in compliance with the EU AI law. So the fact that only Anthropic is doing this so far — Neville Hobson: Undoubtedly. Shel Holtz: — doesn’t mean that you won’t see it in ChatGPT and Gemini and even maybe Grok. We’ll see. Neville Hobson: Yeah. I mean, they talk about retrofitting this to earlier versions of Claude, and that makes total sense to me. You are right. I believe it would make no sense if Claude was the only one doing this. So we’ll expect news from the others. But in the meantime, my suggestion to everyone is: Don’t worry about this. Be true to yourself. Read Clay’s guidelines — what are they called? I can’t remember. It’s kind of a policy, writing policy. Shel Holtz: Yeah, policy. Writing policy. AI writing policy, yeah. Neville Hobson: Okay. And there’ll be others adding to this and saying, “Here’s my version,” and so forth. So you can pick what’s best. But just focus on your writing. One of the points that comes out of this is absolutely right: You do the writing. Don’t just say, “Give me a 1,200-word article on whatever topic.” That’s the best approach, and that has been the case since before this topic emerged. Shel Holtz: And that will be a 30 for this episode of For Immediate Release. The post FIR #526: Forget Anthropic’s AI Watermark. Can You Defend Every Sentence? appeared first on FIR Podcast Network.

22 MIN

AUG 13, 2026

FIR #525: The Consequences Businesses Will Face from Work Slop

We have been hearing about work slop for a while now, and evidence is mounting that it can create long-term problems for companies. Yet we are hearing very little (if anything) about what businesses are doing about it. There’s new research, which Neville and Shel discuss in this episode, along with some thoughts about actions companies can take. Links from this episode: Why People Create AI “Workslop”—and How to Stop It Navigating AI in the Workplace: 2026 Harvard Business Review warns AI ‘workslop’ is rotting companies from the inside Botsitting, botshitting, and the hidden human labor of AI at work 27 Seconds to Breach: The AI “Workslop” Tax Hitting 2026 Bottom Lines How to Eradicate AI ‘Workslop’ From Your Office The next monthly, long-form episode of FIR is tentatively scheduled to drop on Monday, August 24. We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected]. Special thanks to Jay Moonah for the opening and closing music. You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog. Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients. Raw Transcript: Shel Holtz Hi, everybody, and welcome to episode number 525 of For Immediate Release. I’m Shel Holtz. Neville Hobson And I’m Neville Hobson. Shel Holtz And Neville, it is great to have you back on the show. You’re well enough to record again. Neville Hobson Yeah, it’s certainly been an interesting few weeks, including the kind of disconnection that I did from just about literally everything with this heat exhaustion that I suffered, which was quite profound and far more serious than I thought, according to the doctor. And then the R&R, as I described it—the rest and recuperation—slotted in nicely with our existing family plans here to spend some time in Novato, Marin County, in California, which we did last week. And as some listeners will know, you and I and the family, in fact, all met together, had lunch in a delightful restaurant in San Francisco last Thursday, and took some pictures, which is what people saw. So I’m back in the UK now. It was a worthwhile visit, and, well, here we are doing FIR again. It feels like déjà vu all over again. That’s great. Shel Holtz It is back to normal. I should tell people that if you’re curious about the wonderful restaurant we ate at in San Francisco, it’s called Waterbar, and my company, Webcor, built the building that it’s in. So it was fun to have you there and to say, “We built this.” Neville Hobson Yeah, it was. I agree. The hospitality was really good in that restaurant. Very nice service, friendly people, excellent food, of course. It’s focused on fish, so if you’re a fish connoisseur, this is the place to go in San Francisco, I would say. Yeah, really nice. Shel Holtz Yeah, and right on the water. You get a view of the Bay Bridge and a good part of San Francisco Bay. So, yeah, it was great to see you in person. It’s been since 2019. Neville Hobson Yeah, a long time. Shel Holtz And here we are looking at each other over Riverside, but that’s not the same. Neville Hobson No, the next best thing, but not the same. Shel Holtz Yeah, being together in real life really matters. But today’s topic is one that we were planning to discuss just before you took ill, so we’re going to return to it. Neville Hobson Yeah. Shel Holtz We’re going to return to it. Neville Hobson We are. And this is based on an article that caught my eye a few weeks ago in the Harvard Business Review. It was actually published back in January, but having returned to it, I think it’s even more relevant than it was then. The title of the article is “Why People Create AI ‘Workslop’—and How to Stop It.” Timely topic, I’d say. The authors use the term “workslop” to describe low-effort, AI-generated work that looks polished on the surface but ends up shifting the real work onto whoever receives it. We’ve probably all seen examples: reports full of generic language, presentations that say a great deal without saying very much, emails that are technically fine but leave you wondering what the sender actually thinks. Since that article appeared, the wider idea of AI slop has become much more prominent. We’re hearing the term applied to the flood of cheap, AI-generated material appearing on social networks, in publishing, marketing, and elsewhere. There’s even research suggesting that people increasingly use “AI slop” as a judgment about authenticity. Something doesn’t necessarily have to be AI-generated to be dismissed as slop; it just has to feel like it. Perhaps the most useful new evidence comes from research by the Society for Human Resource Management, SHRM, the world’s largest professional association dedicated to the practice of human resources management. Published in June, SHRM’s Navigating AI in the Workplace: 2026, based on more than 5,000 U.S. workers, found that 41% use AI at work, and 44% of those users describe at least some of their output as AI slop. Early-career employees feel the greatest pressure to adopt AI. SHRM also found greater engagement and commitment where organizations take an open approach to AI integration. What interested me most about the HBR article was its argument about why workslop happens inside organizations. The authors suggest that it isn’t fundamentally an AI problem at all. It’s a leadership problem. Many organizations are telling employees simply to use AI without defining what success looks like, without giving people sufficient guidance or confidence, and without creating an environment where it’s safe to experiment, ask questions, or admit uncertainty. In that environment, people can end up optimizing for appearing to use AI rather than actually using it well. Putting it simply, the article says it’s the result of unclear AI mandates and overwhelmed teams. Leaders are issuing vague directives for employees to start using extremely powerful tools, while many of those employees are overburdened, psychologically depleted, and operating in environments where it doesn’t feel safe to admit uncertainty or ask for help. Evidence published since the HBR article seems to reinforce rather than undermine that argument. We’re seeing more research showing significant numbers of employees encountering poor-quality, AI-generated work while, at the same time, employees are under increasing pressure to demonstrate that they’re using AI. There’s another dimension to this that I think is particularly interesting: productivity. Imagine that something that previously took me two hours to produce now takes me 20 minutes with AI. That’s a significant productivity gain for me. But suppose I send it to Shel—to you—and you then spend an hour trying to understand what I mean, checking my claims, and correcting some mistakes that I hadn’t spotted. Have we actually improved productivity, or have I simply transferred the cost of my work to Shel? That’s essentially what workslop does. And some recent research into AI-assisted work describes this as a kind of tragedy of the commons. Individuals can appear more productive while imposing additional costs on everybody else. If an organization measures AI adoption, output volume, or individual efficiency rather than the performance of the whole system, workslop might actually look like successful AI transformation. That struck a chord because it echoes something we’ve discussed on FIR more than once. Whenever we hear stories about AI failing in organizations, the technology itself often isn’t the real issue. AI has a remarkable ability to expose weaknesses that were already there: unclear leadership, poor communication, lack of trust, unrealistic expectations, and organizations measuring activity rather than outcomes. And there’s an important question of individual responsibility here, too. Poor leadership may create the conditions for workslop, but that doesn’t absolve the person who sends it. If I knowingly send something I haven’t properly considered or checked simply because an AI produced it, that’s also a failure of professional judgment. So perhaps workslop isn’t really the diagnosis at all. Perhaps it’s a symptom. Is AI creating a new workplace problem, or is it simply shining a very bright light on old management problems? What happens when leaders measure AI adoption instead of better decisions, better collaboration, and better business outcomes? And as AI makes producing apparently polished work almost effortless, does human judgment actually become more important rather than less? The Harvard Business Review article may now be seven months old, but the questions it raises are becoming more urgent, not less. Shel? Shel Holtz Without question. And there— Neville Hobson Mm. Shel Holtz —there is newer research. I found work done by the Work AI Institute. They did a 2026 research project called The Work AI Index. And in addition to workslop, they have a new term in there called “botshitting.” I kid you not. Neville Hobson Ha ha ha. Shel Holtz Botshitting describes employees knowingly shipping AI output they believe is wrong. And 12% of employees who were surveyed for this research admit to doing this. Interestingly, when they are basically caught, they blame AI. They say, “The AI did this.” Forty percent of workers blame the AI rather than themselves. There was also a study in January 2026 from Workday. Some 3,200 leaders found that 37% of AI productivity gains are immediately lost to rework because the workslop isn’t adequate, with employees spending an average of six hours a week correcting or rewriting flawed AI content. Built In reported workslop now affects roughly 40% of employees and costs about $186 per worker per month. And there was a more recent study—BetterUp sourced this figure—that found that 66% of workers are spending six-plus hours every week fixing AI errors. So this is serious stuff. And by the— Neville Hobson Mm. Shel Holtz —way, nearly every recent report on workslop ties this to layoffs associated with AI. Workers and outlets are connecting 2026 tech job cuts to premature AI-driven headcount reduction. And the workers who are left behind are absorbing that workslop rework burden. So we cut people to use AI; now the survivors are— Neville Hobson Mm. Shel Holtz —cleaning up after it. That’s an important framing to look at. And one last thing I’ll mention from some research—this was just from June of this year. They found that the errors in workslop go beyond the work that somebody else needs to do to correct it. It moves downstream through teams and into the organization’s collective knowledge base. And that knowledge base just deteriorates. So it’s not just individual wasted hours, but this slow erosion of what the organization actually knows to be true. This is a very serious problem, and I don’t see a lot being said about measures that organizations are taking to address it. I think some measures need to be taken pretty soon, or this is going to get out of control. Neville Hobson Yeah. The Harvard Business Review article does have some suggestions about what organizations can do to address it, which I’ll mention in a minute. But they also include some examples. I found these rather interesting. Not a whole lot—three only—but nevertheless, they talk about the toxic effect workslop can have on workplace dynamics, breeding mistrust and leading team members to think less of the sender’s intelligence and trustworthiness, among other traits. They talk about their ongoing research, where they’ve heard a number of examples of workslop seeding ill will, eroding trust, and generally having a corrosive effect on workplace morale. There’s one where I thought, I can imagine this. One example they give is an employee at a technology company who told the Harvard Business Review that he’d noticed the tone in his performance review was unlike his manager, and that the document recycled content from his self-evaluation. The experience made him feel unvalued and underappreciated, and he gave up all hope that he would ever be promoted. So, reading between the lines, what happened? Whoever did that grabbed some of the employee’s self-evaluation and had an AI produce the review. I bet that’s what happened there. We don’t know the details behind that, but that’s a really good example of what you could see happening in an environment where there is a lack of clarity, all those things the Harvard Business Review points out, and the pressure to deliver in some form or another. So, for instance, doing an employee evaluation would come into that area where the manager has cut corners and used AI to do it. And I could imagine that is happening a lot. And in relation to the examples that the Harvard Business Review piece includes, they say something quite interesting. Many of the responses focused on the productivity costs—the time people wasted dealing with each instance of workslop that crossed their desks. But they note that what should really worry leaders is the impact workslop can have on human relationships. And I think that’s a point to hammer home. It’s a kind of soft thing. You don’t necessarily see it, but it’s there. And some of these examples make that very clear indeed. So that’s a manifestation of the management failure behind all of this, Shel, don’t you think? Shel Holtz I do. And I think it’s because we’ve been rushing headlong into this rather than pausing to strategize it. And I think we’ve talked about that deterioration of human relationships once before. The situation is that people are not picking up the phone or sending an email off to the local subject-matter expert because Claude has the answer. ChatGPT has the answer. So you’re getting less of that interaction within the organization. This is causing people to assume some of the work that is outside their area of expertise, outside their lane, if you will, instead of reaching out to the people in the organization who actually have that expertise so they can do the work that previously was part of their job. Another interesting thing you noted was that somebody felt this was leading to a lack of promotion opportunity. Neville Hobson Mm. Shel Holtz I’m reading a book right now, an excellent book. We were talking about it before we started recording. It’s called Robot-Proof. And one of the things she talks about is this economic phenomenon of deprofessionalization, where AI takes on a certain amount of the work, and therefore you, as the professional, don’t need to know or do as much as you did before. And it lowers your overall value. It pushes the wage down for people who are doing that. The author, Vivienne Ming, uses an analogy of Jiffy Lube, which in the U.S. is a place where people take their cars to have their oil changed. And she says, “What if colonoscopies worked that way?” You don’t need a doctor to do a lot of this work because a lot of the work in a colonoscopy is pretty routine and AI can do that. Now you need just the Jiffy Lube guy there to take the hose and do— Neville Hobson Yeah. Shel Holtz —the insertion, looking at the camera to make sure it’s right. But that’s more of a technical skill than a deep medical skill that is taught in medical school and then honed over years of experience and practice. And this is the road I think we’re headed down if we don’t take some steps to fix it. The people who were brought into an organization based on education and years of experience and subject-matter expertise, and perhaps even thought leadership, no longer need all that stuff. They just need to do the base-level effort that is required by a human following the instructions of the AI. So I think organizations need to look at this very seriously: how this work gets done, what is required of the outputs, and having guardrails in place to make sure that people aren’t skirting around it. Because let’s face it, if the cost of producing this stuff continues to drop, there’s an incentive on the part of the organization to go ahead and use this because the cost is lower. It’s hard to see the impact of that on sales and reputation in the short term. In the long term, I think it absolutely will have that impact. Neville Hobson Yeah, I’d say you’re right. I mean, one thing that I’ve been talking about all year in everything I’ve written or said about artificial intelligence in organizations is that it’s about the people, not the technology. If we’re implementing AI or doing a rollout or planning it, or telling people, “Use AI,” as a leader, you have the responsibility to make that as crystal clear as possible and focus on the people. So one of the things I did like about the Harvard piece was the concluding remarks they make, starting with this: “The greatest irony of all is, to make AI work at work, we need to get better at being human.” Absolutely spot on. Leaders need to make space for the unpolished, slower but more rewarding work of human collaboration. Without organizational changes that enable agency and trust rather than AI mandates for overburdened teams, we’ll all drown in the sludge of workslop. A very well-put concluding statement, I think. But the serious element of that is something that, in my opinion, I’m amazed that leaders don’t seem to get at all. They talk about efficiencies. They talk about cost savings. They talk about all these things, and yet there’s nothing about: How do we help employees become more productive themselves and develop themselves, too? How do we help them, in a sense, augment their skills using AI? Now, that’s not to say—and I’m not suggesting for a second—that AI leaders generally just don’t think about that at all. I believe they do, but it’s way down the priority list in terms of how they communicate this. They need to get that to the top of the priority list. But more than just talk about it, they need to put these things in place. And there are many things they can do. There’s, in a sense, rebuilding trust with people. We’ve talked about examples in recent episodes of layoffs that are very clear to see. People are talking about them all the time, and that is happening. We talked about a really great one a few episodes back about what Ford Motor Company is doing with the graybeards. Lovely name. We’ll pass on the age— Shel Holtz Mm-hmm. Neville Hobson —thing on that. But it’s about the organizational collective memory people have of how things are done and what worked in the past, among people who were laid off and are gone. So they rehired a whole bunch, and they’ve benefited hugely, as the articles we cited go into in some detail, including, I believe, if I recall, Shel, they were the number-one automaker for a particular reason that they attribute directly to rehiring all these graybeards. So that is a great way of doing it. But it’s just one example. There are undoubtedly more fundamental things organizationally that people can do to rebuild this connection between the organization and the people who make up that organization—the employees, mostly. So there’s a lot of work to be done. And yet none of this is rocket science to leaders at all, I don’t believe. Just because we’ve got this newfangled technology on the scene, you’re telling people to start using it without giving them the proper guidance. So these, to me, are Leadership 101 things. And if no one’s doing this, then communicators, let’s step up to the plate and help. That’s what I think we should be doing. Shel Holtz Yeah, absolutely. And I have been reading more and more research that talks about what happens when you delegate your thinking to AI. Rather than doing the research yourself and the writing yourself, which helps you think through the issue, it actually has an impact on your ability to think and to learn. So there are a lot of reasons I think that we need to take this seriously. I think the first thing communicators need to do inside their organizations is raise the alarm. Because I think this may be an issue where everybody’s heard “slop” and everybody’s heard “workslop,” but I don’t know that everybody in the organization—especially the decision-makers—is up to speed on what the research is saying about all of this and the deleterious impact it can have on our organizations and the outputs that we produce. So, yeah, I think sharing this data—and we will have links to the research in the show notes—and making the people who make the decisions aware of what this could mean to the future reputation and earnings of the organization is the place to start. Neville Hobson Yeah, I agree. So, simply, there are good tips in this discussion. The links to the articles—please read them. They’re really, really useful. You’ll find a lot of insights in all the ones that we’ll have in the show notes. So there you’ve got your route map, I would say. Shel Holtz And that’ll be a 30 for this episode of For Immediate Release. The post FIR #525: The Consequences Businesses Will Face from Work Slop appeared first on FIR Podcast Network.

22 MIN

Details

Recent Episodes