A shorter update this week: seven presentations and workshops in ten days have left me genuinely tired, but they’ve also been the best kind of primary research about how people are encountering AI day-to-day. One theme came up again and again: how to use AI in content and marketing without losing your audience’s trust. That throughline runs through a lot of what follows.

There are two big themes this week: what new AI agents are making possible, and the need for a more nuanced discussion of AI’s impact on authors and publishers. In particular, Matt Webb’s first-person reflection on agents will resonate with any publisher shipping products internationally in 2026.

This week’s newsletter focuses on the economics of AI becoming more visible. Sometimes that’s positive for publishers: a new tool for estimating the value of licensing your content. Other perspectives are more challenging, particularly on the real cost of AI usage. For publishers, the lesson is that AI strategy must now include cost, value and platform-risk assumptions.

It’s been a week of jet lag, catch-up and presentations, with a workshop, a webinar for PW and my friends at Westchester Publishing Services, and the IPG Summer Summit. Before I shut down for the bank holiday weekend in the UK, some updates on what happened in AI and publishing this week:

Greetings from Portland, where I’m writing this on the Thursday of IBPA Publishing University—or, as my body clock is telling me, about 04:30 London time on Friday… It’s been a long day, but lots of good conversations with publishers have helped to sharpen my thoughts on this week’s news.

I’ve been on the road again this week, travelling to Portland, Oregon for the Independent Book Publishers Association Publishing University, where I’m speaking today. One quick observation, informed by email exchanges with my friend and his AI agent while I was travelling, and several conversations at the convention.

It was a quieter week for product releases, but there were significant legal, policy and research developments. I am in Portland, Oregon next week for IBPA Publishing University: I look forward to seeing some of you there. In the meantime, have a good weekend.

Earlier today I attended an online seminar hosted by the Creative Industries Policy and Evidence Centre, where Dr Paul Crosby from Macquarie University presented recent research on authors and AI carried out in Australia. Many of the sentiments expressed in the research were familiar: author concerns about the morality of AI training and the impact of AI outputs on their livelihood echoed many of the findings of the Cambridge research on AI and the novel published last November. What I found most interesting about the Australian research is that it was interpreted through economics and econometric analysis. One of the key questions was whether, if AI companies train on human creative work without compensation, the long-run incentives for cultural production are weakened. The concern was that more economically marginal forms of publishing such as literary, niche and experimental writing could be hardest hit. Framing the impact on authors and the creative industries as an externality is a helpful way of emphasising to policy makers that the economic benefits of training on copyrighted work accrue overwhelmingly to tech companies in the US, while the costs are potentially distributed across the wider cultural ecosystem globally. This is a textbook setup for undersupply of a public good (creative works).

Happy Friday. I’m catching up after a week on the road and there is a lot to process. This week my reading and thinking clustered around two themes: what happens when the cost of producing content falls, and whether that cost will be stable over time.

I’m writing this on the way home from Chicago, where I gave a keynote on AI and publishing at the ECPA Leadership Summit. As I said in my presentation, this is not an area of publishing I’ve worked in, and my role was to provide an outside perspective—less about the specifics of the sector, more about what AI might look like when viewed from beyond it. It was a really rewarding and thought-provoking experience, and I am very grateful to Jeff Crosby and his colleagues at the ECPA for their hospitality.

It’s been a big couple of weeks for AI image generation and design tools: with major updates from Canva, Google and Anthropic, users are spoiled for choice. The new tools speak to a trend I’m increasingly seeing—and write about below: the baseline for AI use is shifting from competent prompting to confidently delegating to agents. The exam question for publishers is getting used to managing systems of agents rather than using individual tools.

It has been a busy week: updating the materials for and delivering a lunch-and-learn session and another full day training course on AI for IPG members, and preparing a keynote presentation on the future of publishing for a conference in Chicago next week. To prepare for those things, I’ve been reading and talking to a lot of interesting people about how they are using AI at the moment. In between, I made a quick dash up to London for my ten year business school reunion—a room full of smart people from different industries who provide a triangulation to the publishing experience. There are three interesting patterns that kept coming up in all of those different contexts.

A slightly shorter newsletter than last time as it’s been a particularly busy week: several days of travel, and submitting the manuscript of my book to my publisher. More on that soon, but first, some perspectives on AI impact.

If you read my piece last year on sizing and plotting the UK publishing industry based on Companies House data, there’s an interesting coda today with the publication of the Office for National Statistics’ updated Standard Industrial Classification (SIC) codes for 2026. If you’re not familiar with what SIC codes are, you’ve probably given up reading by now—but the short version is that SIC codes are short numbers that describe what business activities companies engage in. Firms choose one or more SIC codes when they register with Companies House.

I don’t set out with a particular theme in mind for each week’s newsletter: sometimes the things I’ve found in my research cluster serendipitously around a particular area, and more often there’s a range of themes. As it happens, this has turned out to be Use Case Week: practical ideas from Anthropic, the BMJ, consultants Fathm, and Paul Ford. If you can’t find something practical to try as a result of this week’s newsletter, you can have your money back.*

Ian Mulvany has a good piece in the latest issue of InPublishing examining challenges and opportunities for publishers. A couple of his points really stuck with me. Under pressure from investors to signal progress, companies are forcing AI into products and services without a clear sense of value. The result is a proliferation of new features that are individually justifiable but risk being collectively incoherent. And while generative AI can speed up the creation of content, code or even products, it doesn’t create additional time—in other words, opportunity cost still applies. I’ve started thinking of this pattern as strategic debt.

I’m sending this week’s newsletter a day early, as Friday and Monday are public holidays here in the UK—judging by the out-of-office messages I’ve seen this morning, it looks as though a lot of publishing has already checked out for the long weekend. But if you’re still at your desk, some reading matter to take you into the Easter holiday: there’s a lot going on, and a surprising number of strategy signals in this week’s news.

I took part in a webinar on AI and publishing earlier today, hosted by the Crius Group, alongside my friends and colleagues Cameron Drew and Simon Mellins. One of the topics that came up was what mistakes we saw organisations making with AI. The conversation was about publishing, but the more we talked, the clearer it became that these failure modes aren’t industry-specific. They show up everywhere. Here are eight of the most common. If you recognise more than two or three of them in your organisation, you may not have an AI problem—you may have a strategy and operating model problem.

For authors, publishers and technology companies, it’s been a week that challenges the old axiom that there’s no such thing as bad publicity. The practical uses highlighted in the rest of this week’s newsletter need to be balanced against good strategy and judgement.

This piece was commissioned by James McConnachie and first published in the Spring 2026 issue of The Author, the journal of the Society of Authors. It is reprinted with their permission.

It’s been a busy week, with significant developments in copyright policy and a steady drumbeat of new research and product launches. The key question posed by new research linked to below has stayed with me: for the right task, AI brings real benefits, but how do we determine what the right task is?

I gave the opening keynote at the ALPSP University Press Redux conference in Liverpool earlier today, discussing how generative AI is affecting scholarly publishing. My professional background is in trade rather than scholarly publishing, so instead of focusing on the university press business model or specific AI tools, I tried to step back and look at a macro question: what happens when AI changes both how knowledge is produced and how it is discovered.

I’m catching up after three intense days at London Book Fair. That intensity has been mirrored in the AI world, with a slew of copyright and licensing developments, some new practical tools for publishers, and catastrophically misjudged AI developments at Grammarly, Amazon and McKinsey.

I’ve just had three very busy days with friends, colleagues and clients at the London Book Fair. I spent time with every company in my portfolio, and had a great discussion with my publisher about my forthcoming book (watch this space for more). While the show is still fresh in the mind, here are a few reflections—less a comprehensive account of the Fair than a set of personal observations, raw signals and conversations that struck me over the last few days.

It has been a week in which AI moved decisively out of product demos and into politics, newsrooms, courts, and war. Across very different domains, the same underlying tensions keep surfacing: control, accountability, and whether speed is crowding out judgement.

A slightly shorter—and later—newsletter than normal, as I’ve been on the road this week (six webinars and courses in four days—welcome to new subscribers from those events). I started the week needing a productivity boost, and Google provided one…

Everything is now a data business, especially media companies. My friend Alex Boden’s analysis of the Washington Post’s pivot to WP Intelligence is characteristically sharp: editorial expertise converted into structured intelligence products, sold to professional audiences on enterprise contracts. The playbook works for a news publisher. The question for trade book publishers is what version of that pivot is available to them.

In recent months there has been a discernible shift in AI—and the pieces I feature—away from simple chatbots to agentic tools like OpenAI’s Codex and Claude Code. My smartest friends have moved from using AI for five or ten-minute microefficiencies to delegating complex coding or research tasks measured in hours. Chatbots will still have their place, but the rise of AI agents and big themes around trust dominate my thinking this week.

Thanks to Luis González, Javier Celaya, Christopher Kenneally and their colleagues for inviting me to give a keynote address on the impact of AI on audiobook publishing this morning at Parix Audio Day 2026 in Madrid. It’s one of the best conferences in the publishing calendar, and hands down the best venue.

I’ve been trying to write something here roughly once a week, largely but not exclusively about my experience working with AI in publishing. I hadn’t planned to post anything else this week, but John Willshire linked to a post from Rachel Andrew on managing publishing operations in the context of generative AI and I wanted to respond to it with some thoughts about how the issues it raises affect book publishing.

Andrew Savikas recently wrote that one of the best ways to see the future of knowledge work is to watch how software developers work. It made me smile because I got to know Andrew nearly twenty years ago through his work at O’Reilly Media and their Tools of Change for Publishing conference—“watching the alpha geeks” was a core philosophical and strategic tenet for O’Reilly.

Only two out of nearly three dozen publishing job ads I reviewed this week mentioned AI skills. Meanwhile, the marketing world has moved on entirely. This issue explores that disconnect, alongside ElevenLabs becoming the UK’s first AI and media decacorn, Amazon’s new publisher data marketplace, and what agentic search tools could mean for editorial archives.

I’m on my way back from the IPG Spring Conference in London, always one of the highlights of the publishing year. It was a really good mix of the inspirational, the practical, and a uniquely generous, welcoming and supportive community. I particularly enjoyed the presentations from former Shazam chairman and crime writer Ajay Chowdhury, National Literacy Trust CEO Jonathan Douglas, BBC media editor Katie Razzall, Illumicrate CEO Daphne Tonge and Bonnier UK Co-CEOs Sarah Benton and Jonathan Perdoni. I spoke in the morning about the IPG’s forthcoming short training courses (data, content creation, websites/GEO and AI risk/governance) and in the afternoon to give members an update on the policy and licensing landscape for AI.

This week’s stories play out at a distance—in financial markets, academia and big tech—but they point clearly toward future directions for publishers. At the other end of the scale, I’ve written about my hands-on experience with Claude Code and agentic AI, which you could start using today.

There are two big themes in this week’s newsletter: utility and accountability, and they don’t sit easily together. The productivity gains many of us are seeing from AI are increasingly in tension with unanswered questions about training and transparency.

This week’s headline stories sit at the uncomfortable boundary between assistance and dependence. From lost intellectual scaffolding to new AI “coworkers”, the question isn’t just what AI tools can do—but how we should use them without dulling the human thinking they’re meant to support.

This week’s AI stories are less about capability and more about control: how models are trained, when licensing works, and where consent is quietly assumed rather than granted. For publishers, these mechanics are starting to matter more than the rhetoric.

Happy New Year. The first AI stories of 2026 aren’t about speed or scale, but about authenticity—who is speaking, what can be trusted, and where responsibility sits when things go wrong. That’s uncomfortable territory for publishers, but increasingly unavoidable.

(Originally posted on LinkedIn): Every December since 2020, I’ve posted about my books of the year: to mark the passing of the year, as suggestions for anyone looking for holiday gifts, and to start a conversation with my network about books at a time when many of us would like to read more, but struggle to (I wouldn’t have got through many of them without the option to listen as audiobooks on xigxag).

Publishing seems to have been slowing down for the holidays for the last couple of weeks, and while AI isn’t on quite the same trajectory, there are fewer notable developments this week. I hope that you have a good festive break if you’re taking one—I will be, so there will be no newsletter next week. Normal service resumes in January.

A slightly shorter newsletter than normal this week, as I have been balancing four days on the road for a conference and board meetings, and trying to close off various projects before the holidays. I’m planning to do one final newsletter for the year next Friday, 19 December, then take a week and resume publication in the New Year.

This was a week in which AI refused to behave according to anyone’s preferred narrative: strong environmental claims unravelled, legal cases tightened, policies were challenged, and AI models themselves showed new points of fragility. I was particularly interested in new research that challenges a common false dichotomy between AI use and creativity. ​ Talking to publishers, the environmental impact of AI is regularly cited as one of the major reasons not to use it—and it is one of the central criticisms in Karen Hao’s bestselling book Empire of AI. But researcher Andy Masley recently highlighted significant methodological errors in Hao’s calculations, which overstate water usage by several orders of magnitude. In response to Masley, Hao has acknowledged the error and will correct the book. The fine detail of this is quite abstruse, but for me the takeaways are simpler. The response to the controversy is a Rorschach blot for how people feel about AI. Many of the people commenting in support of Masley condemned Hao for being ideologically anti-AI. In turn, some of Hao’s supporters dismissed Masley out of hand for his links to the Effective Altruism movement. Reality is more nuanced than either characterisation: Hao may be directionally right even if some claims are discredited, while Masley’s analysis highlights an important epistemic risk, namely, how easy it is to take weak analysis at face value when it aligns with a preconceived worldview. ​ Second, the uncertainty over environmental impact exists in large part because the standard of sustainability reporting from AI companies is so poor—it was interesting to consider this in the context of Satya Nadella’s comments this week that tech companies need to earn social permission for their resource use, a standard the AI sector is currently failing to meet. ​ A fundraising document for the AI music platform Suno claims it is generating the equivalent of Spotify’s entire catalogue every two weeks. I’d love to know the equivalent number for generation of self-published books; Amazon/KDP is probably the only organisation with data robust enough to attempt that calculation, and they don’t publish it. It’s also worth noting that music’s subscription economics make this sort of volume easier to monetise than in books, which remain largely sold a la carte. So if this isn’t happening at this scale in publishing yet, it’s not because it would be technically difficult—it’s because the incentives haven’t lined up. ​ The Information (paywalled) reported that OpenAI is forecasting 220 million paid subscribers to ChatGPT by 2030, which it compares favourably to Spotify’s 280 million paid users. I’d suggest a better comparison is with Microsoft Office, which has just shy of 500 million paying users: to get to just under half of that in less than a decade would imply a remarkable level of ubiquity. Whether the forecast is achievable, or enough to justify OpenAI’s stratospheric valuation, is another matter. ​ However, not everything is going OpenAI’s way: a court ordered that internal communications related to the company’s use of pirated books for training, which OpenAI argued were legally privileged, should be revealed to authors suing the company. OpenAI’s in-house legal team will now be deposed. As I noted after earlier disclosure losses, there is a long road to resolution of this litigation. But Bartz v. Anthropic set an important precedent: a court suggested that training on lawfully acquired books may qualify as fair use, but obtaining books from pirated libraries may be copyright infringement. If OpenAI’s disclosures reveal extensive use of illicit material, pressure for a settlement would likely increase. ​ A major new analysis of 168 million UK job ads finds that demand for AI skills and creativity skills is rising together, especially since the release of ChatGPT, with the strongest overlap appearing in highly-skilled roles and the UK’s established creative clusters. For publishers, this reinforces a point many of us are already seeing in practice: AI skills are becoming a complement to good creative judgement, not a replacement. It also suggests that publishers who invest in both sets of skills, rather than treating them as competing priorities, will be better positioned as GenAI becomes embedded across publishing workflows. ​ Epic Games CEO Tim Sweeney started a debate in the games business by arguing that games marketplaces like Steam shouldn’t label games made using AI, because it will soon be part of how all games are made. Consumer signalling is an equally hot issue in books, though no-one is asking for less transparency, with at least four organisations offering an AI-free checkmark for book covers. ​ Mistral launched a new range of models. On a like-for-like basis, they are not really competitive with larger models from OpenAI, Google or Anthropic, but the interesting aspect is some of the smaller models, designed to be run locally on smartphones or laptops without an active Internet connection, which reportedly benchmark in line with competitor models four times their size. ​ Researchers from MIT, Northeastern and Meta have shown that some LLMs can over-rely on sentence structure rather than meaning, answering questions based on familiar grammatical “shapes” even when the words are nonsense. For book publishers, this helps explain why models sometimes hallucinate confidently in unfamiliar contexts, misinterpret bibliographic or rights information, or respond erratically to out-of-domain material. It also highlights a potential safety and quality risk when AI is used in workflows such as metadata creation, research, fact-checking or customer-facing discovery tools, where subtle shifts in phrasing can produce unreliable results.

Happy Black Friday. This Sunday marks three years since ChatGPT first appeared: a toddler in human years, already learning to run, leaving a mess in its wake, and showing signs of what it might grow into. It has even added a new word to our vocabularies—slop—and several of this week’s links explore whether AI-generated content has any value or is just that. ​ ChatGPT introduced a new shopping research feature this week, which matches products to a user query. I tried using it to find some gifts and it did a pretty good job of matching recently published books to recipients, though it pulled pricing and availability from a range of bookshops and publishers. It’s also unclear how frequently this data is refreshed or how well it handles backlist titles. These are questions anyone who cares about book discoverability should be thinking about too. ​ The new 2025 Edelman Trust Barometer on Trust and AI has some really interesting conclusions for anyone implementing AI in the workplace: about 60% of employees would accept AI aimed at productivity rather than cost saving, and the provision of high quality training increased employee willingness to use AI (unsurprisingly, I endorse that message). ​ Google Labs launched a new pedagogical experiment called Learn Your Way, which uses AI to transform linear textbooks into interactive learning materials. They claim an 11% improvement in retention scores for the AI texts over traditional ebooks. There’s a waitlist signup for users to upload their own PDFs: for publishers, that might be an experiment to consider, but it will also be interesting to see what copyright guardrails are built in to control use of third-party content. It also raises rights questions: at what point does an AI-modified textbook become a derivative work, and who owns those adaptive outputs? ​ On the subject of education, EdTechnical is running a forecasting competition on the impact of AI on education to the end of 2028—the same timeframe forward as from ChatGPT’s release to now. I’m sure plenty of subscribers have a view on this, and there are prizes for the best contributions. ​ Rahim Hirji’s newsletter has a fantastic list of nearly sixty AI habits that would get you sued, fired or embarrassed. Based on your score, you can determine whether you’re a cautious sceptic, a normal human or a walking liability according to Rahim’s scale. There’s a part of me that thinks that with such a fast-moving technology, if you’re not experimenting—and occasionally making mistakes—you’re not learning. But it’s certainly safer to learn from other people’s errors. ​ The AI platform Descript has published a useful guide to slop-free content creation, in particular repurposing an existing asset into different media formats. This is something content and marketing teams in publishers do all the time, and the guide provides some clear, practical advice. ​ New research suggests that AI-produced adverts can’t be dismissed as slop, as they perform considerably better than traditional commercial messages. I have questions about the research methodology, particularly the sample size. And the study measures perceived effectiveness rather than real-world conversion—still, it’s a sign that AI-generated creative may not be as disposable as many assume. ​ On the subject of AI slop, The Wrap returns to a subject that I’ve discussed before: the AI podcast studio Inception Point, now generating 3,000 podcast episodes a week, with a team of eight people. It’s easy to dismiss this as slop, and as I’ve previously argued, it hurts the signal to noise ratio for traditional podcast publishers. But it’s working, as Inception Point has over 400,000 subscribers. The piece goes into new detail about how the company operates. What’s striking is how tightly their production model is tied to algorithmic opportunity—filling keyword gaps with astonishing speed. It’s not hard to imagine the same volume x velocity playbook applied to ebooks. ​ A more traditional publisher, The Atlantic, has struggled with AI platforms crawling its site for data: one company tried to crawl over half a million times in a week. This piece looks at its strategy for managing access to its content. What’s particularly useful is the data-led approach that the Atlantic took, using its logs to determine which bots brought referral traffic and which should be blocked (less than a third brought any value). On that point, I’ve spoken to two publishers in the last six months who were proposing to make decisions about their websites without even reviewing their logs. Be more Atlantic. ​ Finally, I don’t think I’ll ever be comfortable listening to my own voice, but I did a very traditional podcast interview with the brilliant Alison Jones this week, talking about our respective careers in books, digital change, and the impact of AI on publishing.

This week has been full of stories that show how fast the conversation around AI, authorship, and creative integrity is moving. From new research on writers’ concerns to real-world disputes over AI-generated artwork, it’s clear that evidence and transparency matter more than ever. ​ A new research report from the Minderoo Centre for Technology & Democracy at the University of Cambridge received significant media attention this week for its headline claim that half of UK novelists believe AI will replace their work (there’s a lot more beyond the headline and I really encourage you to read it, especially if you’re a trade fiction publisher). It’s a strong piece of exploratory, qualitative research, though its focus on novelists alone excludes the equally important perspective of non-fiction writers (particularly since many authors work across both fiction and non-fiction). I have expressed some doubts about the methodology and whether one can reasonably draw population-level inferences from a convenience sample of 258 authors. But this is really to say that I hope it demonstrates the topic is important enough to do a larger scale study across types of authorship. ​ Coincidentally there’s a very useful piece by Marc Zao-Sanders in HBR on how to make sense of research on how people use AI, which makes important points about vested interests and inherent bias. For clarity, I’m not suggesting these are problems with the Cambridge research, only that the principles are worth keeping in mind with every data source on the subject (including what I write). I think I’m going to take this sentence from Marc’s piece as the mission statement for this newsletter: “The path to a sensible, defensible, and useful view of what’s going on lies in the synthesis of many different sources.” ​ Second only to the Cambridge research in press coverage this week was the news that two eminent authors were disqualified from a leading literary prize in New Zealand for the use of AI in their cover artwork, after new rules were instituted by organisers. This highlights a number of issues: the fact that the authors were unaware of the use of AI by their publisher, the concomitant need for transparency between authors and publishers, and the practicality of applying rules on this fairly and consistently. ​ On this subject, I came across an interesting practical AI feature this week: Google Gemini now has the ability to look at an uploaded photo and detect SynthID watermarks that are added by Google’s own image generator to determine whether it is likely to be real or generated. It takes one to know one? This could become a helpful tool for publishers trying to verify the provenance of submitted images. ​ However good automated detection gets, it’s never going to be completely foolproof in detecting AI, and there’s a real problem with false positives. How do you defend yourself if you’re accused in error of using AI? (This is far from a theoretical problem: I know book publishers this has happened to in the last year.) This is a really interesting case study from the brilliant Watershed in Bristol, which was accused of using AI in its marketing. Their response is really clear, and includes an explanation from their designer. It’s a model of clarity, and it raises an uncomfortable question: how many of us could offer an equally confident and well-documented rebuttal of our creative processes? ​ I suspect Google Scholar is one of the company’s lesser-known offerings, but it’s an essential tool for many researchers and academic publishers. This week Google released an updated, AI-powered search called Scholar Labs in limited preview. This is particularly helpful in providing a short summary of relevance to a search topic for each item returned, and could be especially valuable in the exploratory stages of research or literature review. ​ I’m always interested in AI case studies from other industries, and this post from broadcasting about developing a complex content workflow in under thirty minutes offers a compelling look at how AI can accelerate production. Of course, the prerequisite for doing this was having an MCP server that already interacted with key systems: the book or journal publishing equivalent would first require an existing integration with bibliographic databases, content management systems and other infrastructure. But the underlying principles hold true. I particularly liked this assessment of the developer’s role in the results: “To be honest, I don’t see a world where AI replaces engineers. It’s more about all engineers operating at a fundamentally different velocity, albeit constrained by purpose. The knowledge I’ve accumulated over three decades didn’t become irrelevant—it became leverage. I knew what to ask for. I could evaluate whether what [AI] produced was sensible… The AI handled the tedious translation of intent into implementation, yet the customer still owns the ‘purpose’ that drives the intent.”

There’s an interesting new report out from the Minderoo Centre for Technology and Democracy at Cambridge University on the impact of AI on the novel, authored by Dr Clementine Collett. There’s much that I completely agree with and that accords with my experience of working with book trade on AI. But I have some constructive thoughts about the framing/communication.

Hello from Dublin, where I’ve been speaking at the Publishing Ireland conference. It’s been a really excellent day and good to see many subscribers and friends here. The big theme this week is AI developments from Big Media: new features and guidelines from Spotify, Netflix and Time all have implications and lessons for publishers of all sizes. ​ Spotify has introduced a new feature called Recaps, which uses AI to generate a short narrative reminder for listeners when they come back to an audiobook after a while. The carefully worded announcement states clearly that book content is not used for LLM training—though it seems likely that it is processed by an LLM for generation, even if the data isn’t retained or used to train the model. Functionally, this is similar to Amazon’s recent Story So Far feature. There are two immediate implications for publishers. With two of the leading retail platforms offering this feature, similar functionality is likely to come to other reading systems. Second, the tight integration of AI into core platform functionality shows how difficult it is becoming to accommodate requests from authors who want to avoid any AI use associated with their work. ​ Generation of summaries is also the focus of this opinion piece from the Authors Alliance, which highlights an implication of the recent court ruling in the OpenAI class action: that certain short summaries of fictional works might, in some circumstances, be interpreted as infringing derivative works. This isn’t an issue for Spotify or Amazon, who work directly with rights holders, but the piece speculates about potential risks for platforms like Wikipedia. ​ AI audio platform ElevenLabs has launched a new marketplace for ‘legendary voices’, synthetic versions of leading figures such as Maya Angelou, Michael Caine, Laurence Olivier and Judy Garland for text-to-speech applications. Most of the featured voices are no longer dead, with relatively few contemporary figures. It will be interesting to see whether more living rights holders choose to participate, given the varying legal frameworks around voice likeness and posthumous rights. ​ Netflix has published a set of guidelines for the use of generative AI in creative applications, which is a useful thought starter for publishers. What’s particularly relevant is the thinking on ethics and how AI works with creative workflows and third party intellectual property. Ideally publishers should have policy documents that are internally facing like this one, and for key external partners like authors (as in the Wiley examples last week). ​ Google has introduced a new retrieval augmented generation (RAG) functionality to the Google Gemini API. Put simply, the new File Search tool allows Gemini to interact with your own repository of files rather than relying solely on its general training data and web search. This is priced aggressively at $0.15 per million tokens (roughly 5-6 non-fiction books). ​ Google has also introduced a series of improvements to one of my favourite AI tools, NotebookLM, including a Deep Research model and better integration with PDFs and other files in Google Drive. This makes it significantly more useful as a research tool. ​ As an example of a really expansive publisher AI strategy, Time magazine has introduced an AI Agent offering interactivity, summaries, translation and audio versions alongside its reporting. ​ Helen King’s PubTech Radar has an interview with Jonathan Woahn, the founder of AI content management platform Cashmere, which sits between publishers and AI companies. It’s all useful context, but the key takeaway for me is that publishers can either engage with licensing and shape how it works, or watch it happen to them.

I sent the first of these newsletters a year ago this week, to twenty-seven friends and colleagues. Since then, it’s grown to over 1,500 readers and has become one of the highlights of my work. The cadence of researching and publishing weekly keeps me up to date before I hit send, and I really value all of the comments and connections that have come from it. Thank you for your time and attention, and I hope you’ll stay with me for another year. ​ Thad McIlroy has an excellent, thought-provoking piece on the publishing industry and technology adoption, using the slow take-up of ONIX 3.0 as a frame to ask, what if publishing is equally dilatory in adopting AI. Thad evidences his argument in part with the absence of AI requirements in current job listings—to back up that argument, I looked at a number of open roles at UK publishers that talk a good game on AI, and even in technology-forward areas such as digital marketing and analytics, there’s not a mention of AI skills… ​ Authors and publishers suing OpenAI have gained access to internal company discussions on the use and deletion of books from LibGen, which could be seen as destruction of evidence. The plaintiffs are a long way from their day in court, let alone damages or an Anthropic-type settlement, but it’s not a stretch to imagine that as an outcome. ​ Getty Images largely lost its UK lawsuit against Stability AI. The court found only limited, historic trademark infringement relating to Getty watermarks and rejected secondary copyright claims. For rights holders, the UK still lacks a definitive legal position on whether training AI models on copyrighted works is infringement—meaning policy developments and future cases will likely shape the landscape more than this ruling. ​ One of the fundamental challenges for publishers in the last year has been balancing practical opportunities from AI with litigation over use of content in training. Litigation requires confrontational positioning, but companies have to be careful not to box themselves into a corner. That said, I’m not sure I’ve seen anything quite as quick-footed as Universal Music announcing the settlement of litigation and a strategic partnership with the opposing party in the same press release… ​ Amazon filed a new lawsuit against AI platform Perplexity for allowing its AI agents to shop on behalf of users (Amazon is, of course, developing similar functionality itself). Perplexity responded with an understated argument that this is a threat to all internet users. For publishers who are grappling with issues of AI agents accessing information, it’s a small foretaste of the bigger arguments when agents are transacting with websites. ​ Separately, on a more practical note, Perplexity published a new set of guidelines on AI in the workplace, built around its own tools but with some useful general advice on prompts and capabilities for any AI user. ​ Slate uses recent discussions between local history publisher Arcadia and its authors as the starting point for a piece on publisher licensing to AI companies. Besides providing general context, it’s also helpful in highlighting the opacity of a lot of licensing deals and other friction points for authors. ​ A week after a series of largely incremental product updates from Adobe, Canva announced a complete overhaul of its product line, now described as a Creative Operating System. AI tools are present throughout, including a proprietary design model, integration with LLMs and AI-powered analytics for marketing campaigns. And Affinity, the InDesign competitor acquired by Canva last year, is being made free in perpetuity (and has been downloaded over a million times in less than a week). Overall, this looks like a really compelling alternative to Creative Suite for publishing design and marketing teams. ​ For academic and educational readers, Sage published a new white paper by Tom Chatfield on AI and the future of pedagogy. It strikes a good balance of highlighting issues and opportunities, particularly the potential for AI to be a guided learning partner, not merely a source of content. Anyone developing courseware and textbooks could get a lot from it. ​ On that theme of AI as partner not replacement, belated thanks to Shimmr founder Nadim Sadek for sending me a copy of his excellent new book, Quiver, don’t Quake, a combination of philosophical musing and practical guidelines on AI and creativity. It made for a thought-provoking return journey from New York last week. ​ John Wiley published a new set of guidelines on the use of AI for journal authors, editors and peer reviewers. This complements their earlier guidelines for book authors, which I’ve often held up as the gold standard for publisher policy and communications. Both sets of guidelines are essential reading and Wiley has done the entire ecosystem a service by making them publicly available. It would be good to see other publishers being similarly proactive (if your publisher has a set of guidelines, let me know and I’ll share it). ​ The Wiley guidelines underpin clarity and transparency on where AI is being used, which seem essential principles. But there’s a tension in this: new research suggests that readers’ perceptions of content shift when they are made aware of AI involvement, with lower levels of trust and engagement—though this is offset when readers have greater levels of AI literacy. To be clear, I do not think anyone should take this as justification to obscure AI use, but it’s a reminder that AI remains a nuanced and contested topic. ​ Overdrive recently added an AI book curation and recommendation feature to its library app Libby: in this interview, CEO Steve Potash discusses the feature, dealing with user scepticism, and whether it can help to arrest the decline in reading.

Greetings from New York City, where I’ve been on a working vacation for the last week. It’s my first time back since Covid, and besides the inevitable changes to the city (I could date the last fifteen years of photos by the maturity of the foliage on the High Line) the thing that’s struck me is the ubiquity of technology. Every billboard in Midtown is for an enterprise AI tool, and while this was primarily family time, I had some good conversations with publishers about their use of AI.

For the first time in a while, I didn’t go to Frankfurt Book Fair and experienced it vicariously through Slack, WhatsApp and press updates. My impression was a lot of conversation about AI, and a lot of publishers talking more than doing. Many years ago I was introduced to Seth Godin. After our mutual friend had explained my role at a major publisher, Godin smiled and replied with one of his catchphrases: “Great, but what have you shipped?” I thought of that moment while reading some of the CEO platitudes from the Fair. So what did I miss? If you shipped something new this week, hit reply and let me know.

It’s been a heavy week with travel, an IPG Lunch and Learn update and interesting discussions at the BISG AI Working Group. But before publishing breaks for the weekend and packs for Frankfurt, some interesting AI developments across audio, author apps and retail… ​ AI audio platform ElevenLabs has introduced a new managed production service combining AI voice reading and human editing and post-production. Audiobook production is one of the launch use cases, and pricing starts at $2 per minute. It’s an interesting middle ground: cheaper than human voice audio, arguably better than out-of-the-box AI voice but still not quite convincing, to my ear at least. I also wonder if offering a managed service around it is also a tacit acknowledgement that tweaking outputs for quality isn’t easy. ​ A couple of weeks ago, I commented on a piece about AI use in education and speculated about which self-help author would be the first to create their own AI coaching app. Thanks to subscriber Nikki Howard from Gill Books for pointing out that Tony Robbins was way ahead of my speculation: for $99 a month you could have access to 24/7 advice from his chatbot. Similarly, this week, hedge fund manager and bestselling author Ray Dalio launched a beta version of an AI advice app using a voice interface. For publishers working with brand authors, there’s an interesting question of whether this kind of AI product becomes a core part of the format mix, or if it’s the equivalent of the iPad apps that were launched with great fanfare in the early 2010s, few of which survive. ​ With the holiday season in sight, Adobe Analytics published new research on how shopper behaviour is being influenced by AI: 53% of respondents had used generative AI for research, 40% for product recommendations and 36% for finding deals. This growing usage explains why ChatGPT is integrating buy links into LLM outputs, and Google recently introduced a payment protocol for AI agents. ​ For publishers, this introduces a trade-off between restricting AI crawlers from their websites to prevent training, and allowing them to encourage visibility of products in outputs. OpenAI and Anthropic use different crawlers for training and servicing user queries, so it’s possible to take a granular approach. But Google and others do not, and most publisher websites that I’ve reviewed take an all-or-nothing approach. ​ Web content being scraped doesn’t just affect publishers and retailers, but also universities. For academic readers, the Authors Alliance has a good piece here looking at AI scraping of institutional repositories (which often contain versions of publisher content) and tensions between restricting AI training and Open Access publishing. ​ Ultimately, across both domains, this will be a judgement call for publishers, but I’m reminded of Tim O’Reilly’s aphorism that obscurity is a greater threat than piracy. ​ At OpenAI’s Dev Day, Sam Altman claimed that ChatGPT now has 800 million weekly active users—a staggering figure that, if accurate, would make it one of the largest digital platforms in the world. For context, Microsoft 365 has around 400 million paid seats. In less than two years, generative AI has gone from experiment to near-universal tool, a shift no other technology has achieved at such speed or scale. ​ For a more independent take on usage statistics, there’s a great new report from the Reuters Institute at Oxford University: 90% brand awareness of leading AI tools, all-time AI use increasing from 40% to 61%, and weekly use of AI tools nearly doubling. For publishers, it’s striking that information seeking is the number one use case, ahead of content creation. ​ There are some interesting points in a recent Econsultancy study on marketing and AI: in particular, 47% of marketers described their knowledge of generative AI as advanced or expert: only 4% then passed a test on their knowledge. ​ Finally, Dan Shipper has an interesting essay here commenting on sharp progress in AI performance but putting it in the context of human intelligence. It’s a timely reminder that the good results that are possible with AI depend on enormous amounts of human framing—not least designing tasks, prompts, and evaluation criteria.

Another week, another flood of AI product announcements from Microsoft, Amazon, OpenAI and Anthropic. But some of this week’s news feels particularly consequential: if you publish books to Kindle or have a direct-to-consumer sales strategy, read on…

The approval of the settlement in the Bartz v. Anthropic litigation overshadows most other developments for publishers this week. But elsewhere the week has seen new research, strong opinions on LLMs and web traffic, and a new creative AI tool that could get traction with creative teams in publishing. Have a good weekend. ​ Big news overnight: Judge Alsup has given preliminary approval to the copyright infringement settlement in Bartz v. Anthropic. There will be more information on the operation of the settlement in coming days, and I’ll be covering this in detail in my next policy update for IPG members. As a reminder, it’s not enough for works to be within LibGen: eligibility for the settlement depends on copyright having been registered within five years of publication and before August 2022 (the date of Anthropic’s infringement). It’s also worth reiterating that the settlement is about copyright infringement, not the subsequent training, which Judge Alsup found to be Fair Use. In considering the sufficiency of the roughly $3,000 per author, the Judge asserted that Anthropic could have paid as little as $1 for a (presumably used) copy of a book and trained from that. ​ I was talking to a client this week about how publishing breaks down into individual and group tasks. Most of the successful AI use cases I’ve seen have been about making individuals more productive—fewer have addressed collaborative working. So I was interested to see the latest experiment from Google Labs: Mixboard, an AI-powered mood board tool. It’s currently only available in the US, so American friends, have at it, while the rest of us wait or VPN. But this looks like a really interesting way for design, marketing and other creative teams to collaborate on visual tasks, and it’s the first AI tool I can imagine seeing on meeting room screens. ​ Useful snapshot data in Google Cloud’s 2025 DORA report on the state of software development: 90% of software developers are using AI, 80% reported increased efficiency, and 59% claimed increases in code quality from using AI tools. ​ Research from the Tony Blair Institute tends to divide opinion. But its latest report on AI usage and attitudes, based on polling by Ipsos, also gives a useful contextual snapshot. Some highlights: more than half of UK adults have used Generative AI tools in the last year, and just under a quarter use Gen AI weekly; two thirds of people who feel confident in their AI skills expect it to help them at work while leaving core responsibilities intact, but only 45% of people with lower confidence are optimistic; and looking at secondary education, 37% of respondents were comfortable with the idea of AI tutors compared to 32% against, with 33% in favour of AI taking on some routine teaching tasks versus 38% against. Lots of undecideds on both, but even small moves in this direction would have far-reaching implications for the sector and for educational publishing. ​ On the education theme, HBR has an interesting report on an experiment using Generative AI for professional training which found significantly better outcomes from personalised AI coaching. It’s interesting but all a bit vague: I’d love to know more about the model they used, as opposed to generic references to a Gen AI Tutor. But it’s an interesting direction. I’m curious to see who is the first self-help author to release an AI coaching tool alongside their books and courses. ​ Continuing the theme of discovery and web traffic from LLMs from last week, my friend Ani Attamian has a great piece in The Bookseller on AI optimisation, including recommendations for tools to monitor performance and visibility. I’ve known Ani since she was working with publishers at Google, and if you’re looking for help in this area, she’d be a great person to talk to. ​ Cloudflare’s annual letter outlines a future where AI agents, not humans, are the main readers of the open web, and traffic-based business models continue to erode. Their proposed fix is to facilitate AI companies paying creators. But as Paul Ford points out in his newsletter, that vision turns writers into prompt-chasing content suppliers. The real lesson for publishers? Paywalls work. In a world of aggressive scraping and vanishing referral traffic, subscription models—not scale—are becoming the most reliable path to sustainability, leverage, and autonomy. ​ Nieman Lab has a piece comparing the AI strategies of the New York Times and Washington Post, which provides some helpful advice on AI deployment for publishers: I was particularly taken by the Post’s policy that while generic information can be processed through enterprise AI tools like ChatGPT, sensitive data can only be used with internally hosted language models. Few publishers will need that level of security, but I suspect if you’re writing about All the President’s Men, sensitive data really is quite sensitive (RIP Robert Redford). ​ Finally, since we’ve established that a non-trivial proportion of you are classics nerds, you might enjoy this tongue-in-cheek suggestion from Ethan Mollick for properly acknowledging the role of AI in work: tagging it with the Latin phrase, Fieri iussit.

A slightly shorter email than usual today, as it’s been a very busy week of travel. I had conversations about AI in some interesting places: a publishing conference, an Oxford college and the House of Lords. The latter two particularly felt a long way from Silicon Valley, but the discussions underlined how much the technology is permeating every aspect of our lives. ​ International readers may have missed President Trump’s state visit to the UK this week, but domestically it’s been hard to ignore. There was a slew of major AI investments being announced, but for publishers the most consequential aspect of the week was confirmation from new Technology Secretary Liz Kendall that no commitments had been made to AI companies on UK copyright and AI training. ​ A couple of major updates on AI and ecommerce. Amazon announced an upgrade to its Seller Assistant, which now includes AI agents for tasks like inventory management. Publishers or retailers using FBA can use the AI to review stock levels, minimise charges, identify product opportunities and plan marketing campaigns. This goes beyond basic automation into generating proactive insights. This is interesting in its own right, but also speaks to Amazon’s bullish strategy on AI: this tooling is for FBA today, but it shows the direction of travel for Vendor Central and other systems. ​ Secondly, on the customer side of the transaction, Google launched Agents Payment Protocol (AP2), a standard way for AI agents to be authorised to make purchases on behalf of consumers. More than a third of UK consumers (more than half in the US) are already using generative AI for research and product discovery. Whether purchasers will be happy with the idea of an AI personal shopper is a big question. ​ I was at the Independent Publishers Guild Autumn Conference on Tuesday, and had the pleasure of chairing a panel with Isaac Jones of the BMJ and Lauren Ingram of Next Big Thing, discussing the impact of AI on websites, and how to optimise website content for AI models. Publishers need to strike a balance between controlling how the content on their website is used by AI models, preserving traffic, and being visible and hyperlinked in AI outputs—or, as I thought about it afterwards, having your click and feeding it. This is a fundamental question given that point above on the importance of AI for product discovery. This is a complex and nuanced topic, but as a start on a checklist for publishers, my takeaways from the session were: there will be a range of perspectives even within a publisher—editorial, marketing and technical stakeholders look at this through their own lens, so coordination is essential; quality of content (including specificity, uniqueness and credibility) is fundamental, which should play into publishers’ strengths; and experimentation is key, including putting yourself in your audience’s shoes and using the platforms they do. Do get in touch if you’d like to discuss what this means for your business. ​ Some highlights from OpenAI’s new research paper on how people are actually using ChatGPT: around 10% of the world’s adult population has used ChatGPT; 18 billion prompts are being sent each week; and non-work use cases (advice, shopping) account for 70% of usage, up from about half a year ago. ​ There’s a sobering report here on librarians’ experience of AI: patrons showing reduced critical thinking, and requesting hallucinated books. ​ Thanks to Thad McIlroy for highlighting an essay on AI in the workplace that argues most organisations are essentially in the business of converting work from one format to another. As it points out, there’s a natural role for AI in this, and the implications it identifies for organisations echo the HBR piece I shared last week on consulting firms. Broadly it rings true. But two nuances strike me. First, AI is very good at summarising—many of the examples in the essay are essentially reductive—but often less capable when the task is to expand short form into long form. Second, what AI produces often loses individuality. In Oxford this week, I spoke with a Mandarin translator about how AI translation can be powerful but flattens linguistic nuance. For some tasks in business and publishing, the output may be good enough. But publishing is a business of meaning-making as well as information—and that’s where the human voice remains indispensable.

This week’s stories point to deeper layers of the AI stack: standards, infrastructure and how organisations are evolving. From licensing to litigation to content production, there are clear signs of where publishing may be heading.

This week’s practical insight focuses on two linked topics: onboarding and oversight of AI models. But there are also more philosophical questions from news publishing and Seth Godin, and details of the most open LLM to date. Have a great weekend. ​ Sometimes it’s the simplest insights that stick. In a recent webinar with Rich Ziade and Paul Ford of AI developer Aboard, Rich said something that stopped me in my tracks: “AI is a new hire.” When you onboard a new team member, you think carefully about their role—who they’ll work with, what systems they’ll need access to, how they’ll get up to speed. In a hybrid workplace, we’ve learned that structure and documentation matter more than ever. But with AI, I’ve seen people open up a general LLM with no context or preparation, and then feel disappointed by the results. It’s the equivalent of throwing a new hire in at the deep end and hoping for the best. ​ Here’s my start on a checklist to avoid that. Have you done each of these things for your AI projects? And let me know if there’s anything you’d add to the list.

The AI and content ecosystem is maturing fast. The big news this week is a settlement in Bartz v. Anthropic with implications for other litigation. But besides the big picture, the newsletter highlights new tools and practical things you can try. Have a good weekend.

AI is changing how people produce, discover and evaluate content. This week’s stories show how fast those shifts are happening. From Amazon’s AI shopping assistant to Grammarly’s reimagining of student writer, publishers face questions about visibility, value and voice.

This edition covers OpenAI’s response to GPT-5 feedback with a partial model selector and a 140,000-word context window, Drew Breunig on how long contexts fail and how to fix them, John Willshire’s seven key questions for working with AI, MIT Press’s survey of 6,000 authors on LLM training, Reddit blocking the Internet Archive’s Wayback Machine, Neil Perkin on using AI for scenario planning, and a personal case study on using AI to size and map the UK publishing industry.

Recently I’ve been doing some research on the size and scope of the UK book and journal publishing industry, based on data from Companies House. This research is a preparatory step towards building a database of companies, highly relevant for my work as an independent consultant and as policy advisor for the Independent Publishers Guild, which represents over 600 book and journal publishers of all sizes. The other trade body, the Publishers Association lists 168 members at the time of writing, particularly larger, corporate publishers—though there is some overlap between the two organisations’ memberships. The UK publishing sector is world leading, representing over £7 billion of revenue, and it represents the bulk of my client base. So it’s worth understanding.

As AI capabilities expand, so does the gap between what’s technically possible and what’s useful and meaningful. Many of the links this week address the same question: where does human perspective add value?

This edition covers fresh data on student AI use and ChatGPT’s new Socratic Study Mode, NotebookLM’s new multilingual audio and video overviews, Adobe’s Firefly-powered Photoshop upgrades, content licensing deals from Gannett/Perplexity and Johns Hopkins University Press with benchmark pricing for academic books, Neil Perkin on where AI still falls short, fresh research on the scale of “Shadow AI” use in workplaces, and Ethan Mollick on the “Bitter Lesson” and outcome-trained AI in messy organisations.

It’s been a significant week of contrasting AI developments on both sides of the Atlantic. In the US, the administration moved swiftly toward a more deregulated environment, signalling its preference for Silicon Valley-friendly policies on AI training and licensing. Europe, meanwhile, saw new releases emphasising transparency, environmental responsibility and public benefit, underlining a growing divergence in approaches and creating important challenges and opportunities for global publishers.

This edition covers ChatGPT’s new Agent mode and what it could automate for publishers, Matt Webb on governance lessons from Anthropic’s vending-machine experiment, a WeTransfer terms-of-service flap, NotebookLM’s new pre-curated notebooks of licensed and public-domain content, Dave Morris on AI in his writing practice, Anthropic’s Claude for Education partnerships, Condé Nast and Hearst signing Amazon Rufus licensing deals, the launch of Latam-GPT, and research on how AI is shaping the words we use.

This edition covers Doug Shapiro’s media-trends slide deck and the questions it poses for any creative business, fresh Cambridge University Press polling on public support for AI training payments, an EU antitrust complaint against Google’s AI Overviews and grim data on news click-throughs, OpenAI and partners providing AI training to 400,000 US teachers, hidden prompts being inserted in academic papers to game AI reviewers, Bloomberg on the “tiny teams” era and Anthropic’s Project Vend experiment as a counterpoint, and research showing managers using AI to make decisions about their direct reports.

Happy Friday—though I’m writing this the day before on the way back from a great day at the Publishers Licensing Services Conference in London. The agenda and other speakers were superb, offering plenty of food for thought. It was also great to meet Helen King, whose PubTech Radar newsletter I’ve really enjoyed recently (do sign up for it!) Connecting with Helen via Bluesky commentary on the conference felt nostalgically like the Twitter backchannel at publishing events in the early 2010s. Thanks to PLS and the IPG for the invitation to speak.

It’s been a really significant week for legal developments: while the newsletter has more of a copyright focus than usual, the courtroom updates are balanced with some really interesting technical developments from Creative Commons, Anthropic and others (skip down if you’re less interested in the legalities). It points to the fact that, however long a road to a settled legal and licensing position, there are immediate practical uses for AI in publishing.

This edition covers Amazon CEO Andy Jassy’s bullish AI update and what billions of agents mean for supplier teams, Turing Institute/LEGO research on how children and teachers are using generative AI, a PLS consultation on licensing for AI training, new image-generation tools from Midjourney and Adobe Firefly, ChatGPT’s new Record mode on Mac, Reddit’s Community Intelligence product, a New Yorker piece on what AI is doing to reading, and the resignation of UK PM AI Adviser Matt Clifford.

It’s Friday 13th. Unlucky for AI image platform Midjourney, which is about to find out why “don’t mess with The Mouse” (or an earthier paraphrase) is a popular aphorism in media law. Luckier for publishers, with some powerful new tools this week: in particular, Google’s Deep Research could be a game changer for production of ancillary content.

Thirty weeks of doing this—thanks for sticking with me. Someone asked this week if this newsletter will be going paid, and to be really clear: NO. I enjoy researching and writing it, and it’s a wonderful way to start conversations. However, if it’s something that you find useful, please do share it with your colleagues. Personal referrals make a huge difference to me, and every one is genuinely appreciated.

Maybe there will be a quiet week for AI and publishing. If so, we haven’t seen it yet. Across a range of stories this week, the clear theme is that while AI might automate the output, the value lies in the inputs: good content, sound metadata, thoughtful contracts and sound human judgement.

This edition covers the US House passing a tax bill with a ten-year moratorium on state-level AI regulation, OpenAI’s continued lobbying for Fair Use, Anthropic’s new Claude 4 hybrid models, the announcements from Google I/O including SynthID Detector, Shopify’s AI enhancements and MCP integration, Anil Dash on Model Context Protocol as a Web 2.0 moment, a new paper on AI in higher education, the Chicago Sun Times’s hallucinated book list, and Steven Bartlett’s fully AI-written-and-voiced podcast.

I’m writing this from a very sunny Edinburgh, where I presented to the lovely Canongate team at their away day. It was great to get off camera and out of the office and have conversations with a group of brilliantly creative people on their event theme, ingenuity, and how AI can support it. If you’re looking for a speaker for your conference or event, do let me know.

This week’s newsletter starts with a question that’s been quietly nagging at many of us working with generative AI: are we saving time, or just skipping the thinking? A brilliant post introduces the idea of “cognitive debt”—the mental version of technical debt—where shortcuts today can cost us clarity tomorrow. It’s a useful lens for publishers figuring out how to scale AI responsibly, especially as new tools promise more speed, but not always more understanding.

There’s a lot to unpack this week, including a couple of quite contrarian views on the environmental impact of AI and on the use of copyrighted material as training data. But I’m starting with something that’s both exciting and very practical…

This edition covers a correction to last week’s link on LLM task length, the new native AI text-analysis formulas in Google Sheets, the latest Adobe Firefly upgrades, eye-catching data on AI-generated music on Deezer and rising submissions to publishers, Wikipedia’s training-optimised Kaggle dataset, an Anthropic paper on how Claude expresses values, Ethan Mollick on “Jagged AGI”, Tom Goodwin’s AI-flavoured update to the Eisenhower Matrix, Trump’s executive order on AI in schools, the new CLA/PLS/ALCS Generative AI Licence, Ahrefs research on AI overviews suppressing click-through, and a contrasting Washington Post deal and Ziff Davis lawsuit involving OpenAI.

This edition covers the headline findings of Stanford HAI’s 2025 AI Index Report, fresh data on the rapidly growing length of tasks LLMs can perform, OpenAI’s o3 model release, the importance of human leadership in AI initiatives, a Bloomberg investigation into Inkitt’s AI-powered romance factory, and an embarrassing case of an AI-generated passage slipping through Springer Nature’s editorial process.

This edition covers Shopify CEO Tobias Lütke’s AI-first hiring memo, a behavioural study suggesting reader preferences for human authorship are weaker than stated, an argument that AI may improve content quality by writing for the model as well as the reader, Anthropic’s launch of Claude for Education with Socratic prompting, a salutary LSE blog on “efficient inefficiency”, OpenAI’s persistent-memory upgrade to ChatGPT and the Temporary Chat workaround, Meta’s Llama 4 multimodal models, and a court ruling allowing the New York Times’s copyright case against OpenAI to proceed.

This edition covers Amazon’s new Nova Act browser-agent model, Google’s Gemini 2.5 reasoning model, Rachel Coldicutt’s responsible-AI dos and don’ts, fresh CLA research showing 82% of UK professionals upload third-party content into AI prompts, FT analysis on the limited employment impact of AI so far, Tyler Cowen on AI in his writing workflow, a Nature piece on AI and academic peer review, and new data points on AI scraping bots overwhelming publisher infrastructure.

This edition covers OpenAI’s inline image generation in 4o and the Studio Ghibli controversy that followed, a US court’s refusal to halt Anthropic’s training on song lyrics and what that means for publisher litigation strategy, a new world map of AI copyright lawsuits, MIT Tech Review on the dangers of total autonomy for AI agents, OUP’s AI Discovery Assistant with Silverchair, the worsening impact of AI crawlers on open access infrastructure and Cloudflare’s response, and a BookBrunch op-ed on AI in bookselling.

This edition covers Meta’s attempt to block Sarah Wynn Williams’s memoir Careless People and The Atlantic’s forensic exposé of Meta’s training data, the billion-download milestone for Llama, attitudes to AI among Ibero-American publishers from a trip to Madrid’s Parix IA conference, the Thaler v. Perlmutter appellate ruling reaffirming that copyright requires human authorship, Y Combinator’s report on AI-generated startup code, Baidu’s aggressively priced Ernie models, Anthropic adding web search to Claude, OpenAI’s updated voice models, Zoom’s new AI agents, and details of a forthcoming Publishing Scotland webinar.

This edition covers OpenAI’s “freedom-focused” pitch to the Trump administration on AI policy, an FOI revealing Peter Kyle’s use of ChatGPT for policy advice, Wiley’s gold-standard author AI guidelines, Adam Hyde’s open-source EasyJournal and the wider rise of AI-assisted “vibe coding”, time spent with Ukrainian publishers and the threat of AI training on Russian propaganda, an HBR piece on responsible AI, Sakana’s AI-generated paper passing peer review, OpenAI’s new Responses API for building agents, and AI’s growing role in commissioning decisions across creative industries.

This edition covers Google’s new AI Mode Search and what it means for publisher referral traffic, Mistral OCR as a low-cost tool for digitising older documents, Canva’s State of Marketing and AI benchmarks, Lisa Nandy’s reassurances on creative-industries support at the Creative UK summit, the Authors Alliance submission to the UK AI and copyright consultation, an argument that fears of AI replacing human writing are overstated, and a note ahead of London Book Fair.

This edition covers a busy week at the IPG Conference and a striking Ethan Mollick chart on AI’s collapsing cost and rising performance, Amazon’s Alexa+ upgrade and what agentic shopping means for book discoverability, an HBR piece on AI agents reshaping consumer behaviour, ElevenLabs’ new ElevenReader Publishing offer for synthetic-voice audiobooks, Claude 3.7 Sonnet’s switchable reasoning mode, the arrival of GPT 4.5, and the UK launch of OpenAI’s Sora video generator.

(Originally published on LinkedIn): Like many people in publishing, I’ve been formulating a response to the UK Government’s Consultation on AI and Copyright. With many leading industry and advocacy groups encouraging contributions, it seems likely there will be a big response. So I was interested to read last week that one of the projects developed by the Government’s Incubator for Artificial Intelligence is a tool to use AI to analyse the results of large public consultations.

This edition covers The Guardian’s licensing deal with OpenAI alongside the New York Times rolling out internal AI tools, reflections from a BookNet Canada Q&A on how publisher questions are shifting from theory to practice, the limits of OpenAI and Perplexity’s Deep Research products, the ElevenLabs/Spotify audiobook partnership against Audible’s stance on AI narration, Descript’s new Custom GPT, and concerns about the UK Government using an AI tool to analyse its own AI and Copyright consultation responses.

This edition covers the fractious end to the Paris AI Action Summit and Mistral’s new Le Chat, OpenAI’s plan to unify its product line into GPT-5, Adobe’s commercially-safe Firefly Video Model, a Thomson Reuters fair-use ruling against an AI competitor and the Authors Alliance critique of it, the rise of AI-enabled “personal software”, Matt Webb on agentic AI and Model Context Protocol, and a practical tip for getting ChatGPT and Claude to write in British English.

This edition covers Adam Hyde’s hands-on demonstration of OpenAI’s Operator agent for research publishing workflows, the unredacted Meta court documents alleging the torrenting of more than 80TB of pirated ebooks, Amazon’s bet on automated reasoning to suppress AI hallucinations, the proliferation of Gemini 2.0 model variants and what model fragmentation means for non-expert users, and a write-up from Jellyfish on building AI agents inside creative agencies.

This edition covers the arrival of DeepSeek and what it means strategically and economically, Apple’s iOS 18.3 changes pausing news summaries and switching Apple Intelligence to opt-out, the wider shift among big tech firms to make AI the default rather than the upgrade, the US Copyright Office’s second report on AI and copyrightability, the UK Government’s AI and Copyright consultation alongside a Government tool that uses LLMs to analyse consultation responses, a mini-literature review on generative AI in pedagogy, and details of the AI-heavy IPG Spring Conference.

This edition covers the $500 billion Stargate AI infrastructure consortium, impending litigation between the News Media Alliance and an unnamed AI company, OpenAI’s new Operator browser agent in the US, Google’s bundling of Gemini into Workspace and the resulting privacy questions, Goldman Sachs research on small business attitudes to AI, a Nature piece on LLM hallucination rates, an AI book-recommendation app surfaced by Storythings, and personal news on an Arvon masterclass for authors and a non-executive director role at Burleigh Dodds.

This edition features stories on the UK’s national AI strategy, a study of LLMs in education, ​Bloomberg editor-in-chief John Micklethwait on AI in journalism, and new features from ChatGPT and Copilot.

This edition covers Fable’s offensive AI-generated reading summaries and the lessons for publishers, Scott Belsky’s “Small Teams, Big Businesses” 2025 prediction, Google Cloud’s roundup of 300+ generative AI use cases, fresh allegations about Meta’s use of pirated books for training, an Oxford Internet Institute paper on AI’s net positive impact on employment, and a hands-on test of Avataar’s Velocity product-video tool.

For anyone taking stock of AI developments as the year turns, Simon Willison’s end-of-year roundup is phenomenally good. I recommend reading it in full as it gives a really comprehensive view of what happened and why it matters, but there are several points highlighted below which are particularly relevant to publishers.

(Originally posted on LinkedIn): Every December since 2020, I’ve posted my favourite books of the year here. It’s the biggest month of the year for book sales, many of us will catch up with reading over the holidays, and with the space for book reviews in the media shrinking, authors need all the shares they can get. Caveat: there’s no fiction here (I just haven’t read enough this year) but if the list is a little wonky in both senses of the word, I can genuinely say I enjoyed every book.

This edition covers the UK government’s consultation on copyright and AI, an Amazon job listing that hints at deeper AI use inside the Kindle reading experience, NotebookLM’s enterprise update, Apple Intelligence’s BBC misattribution, a US AI litigation tracker, and what the Betty Crocker Effect can teach us about generative AI.

This edition rounds up a packed week of tech news—OpenAI’s Sora rollout, Google’s Gemini 2 and Deep Research, and Reddit’s new AI search tool—alongside new public-domain training datasets, AI deal trackers, the impact of AI clauses in publishing contracts, and a four-part model of how organisations use AI.

This edition examines the ALCS report on authors and AI training, Pleias’s openly-licensed and energy-efficient LLMs, new content deals from OpenAI/Future Plc and Bertelsmann/ElevenLabs, the risks of AI-generated podcasts, and practical ideas for working creatively with AI.

This edition looks at Steven Johnson’s Long Context project, the Activate 2025 outlook on AI search and disintermediation risks for publishers, the rise of AI shopping assistants like Amazon Rufus, and tech news including Amazon’s expanded Anthropic investment and the new Model Context Protocol standard.

This edition unpacks the HarperCollins–Microsoft AI licensing deal and its wider implications for trade publishing, OpenAI’s o1 models reaching the API, Gemini’s new memory feature, Ben Affleck on AI and craft, research on AI writing in Substack newsletters, and US dates for the IPG AI training course.

This edition covers Google’s new AI video tool for Workspace customers, a US federal ruling dismissing a copyright suit against OpenAI, fresh data on AI adoption in publishing jobs, AI’s role in research integrity and creative writing, and award news for the IPG AI training programme.

This edition covers the dispute over an AI-generated cover for Jodi Picoult’s new book, Anthropic’s new CSV analysis feature in Claude, the launch of ChatGPT Search, and how AI-driven search could reshape publishers’ relationships with the open web.

(Originally posted on LinkedIn): It’s become a habit each December for me to post about the best books that I read this year. For the fourth year, a selection of non-fiction books that I hope you might find interesting.

I don’t use TikTok, but it has been hard to escape its impact on the publishing industry, where I spent the first part of my career. Nielsen research suggests that one in four UK book buyers used TikTok in 2022, accounting for 90 million purchases. As a collective phenomenon, BookTok was awarded the person of the year award at last year’s Futurebook conference. In the meantime, there’s little doubt that the platform has a huge impact on commissioning, marketing and selling books, and that publishers are investing time and money to make the most of that. As a proxy for the level of interest, a search on the Bookseller magazine’s website this morning showed eighty nine search results for ‘TikTok’ for stories published in the first quarter of the year, versus four results for ‘GPT’. (Long term, I know which one I would bet on being most significant and transformative, but that’s another story.)

This piece was commissioned by James McConnachie and first published in the Autumn 2021 issue of The Author, the journal of the Society of Authors. It is reprinted with their permission.