Context Window 72
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.
AI provokes strong reactions: many individual publishers I speak to, and some organisations, raise serious concerns about its environmental, social and economic impact. But research published this week goes further: nearly thirty percent of workers surveyed admitted to sabotaging their company’s AI strategy—refusing to use tools, ignoring policy, even deliberately uploading company documents to insecure AI models.
This resistance to AI is particularly concentrated among Gen Z and younger employees, motivated in large part by fear of the impact of AI on jobs. Interestingly, that generational perspective was brought up separately this week by Abrams CEO Mary McAveney in comments at Bologna Book Fair.
But it suggests uncomfortable questions for both employees and managers. What’s the greater threat to someone’s employment security, engaging with the technology or being caught deliberately undermining company strategy? And if organisations fail to deliver ROI from AI experiments, should they look to technology capability, problem framing or employee compliance?
For context on AI’s impacts, Stanford’s Institute for Human-Centered AI released its annual benchmark research, and it’s fascinating as ever. If you make time for one long-form report this year, this is the one to go for. It shows AI use accelerating (88% of organisations, 80% of students—lagging the UK figure from HEPI, incidentally). In terms of performance, it highlights the gap between an AI winning the gold medal at the International Mathematical Olympiad, and having a 50% accuracy rate at telling the time from a clock face…
One area where the impact of AI is very measurable: new software, driven by vibe coding of applications using AI coding tools. The Information reported that Apple’s App Store saw a 30% increase in new apps for 2025 versus the previous year; quarterly numbers were even higher, with the first quarter of 2026 up 86% compared to Q1/25.
The Creative Industries Policy and Evidence Centre has a couple of upcoming webinars that look very relevant to publishers thinking about AI impacts: on 7 May, they are presenting research from a study of more than 400 authors in Australia on AI and the economics of creativity, and on 18 June, researchers will present an economic study on how AI effects quality and quantity of new book releases. I’ll be attending both and will write about them in subsequent weeks.
The publishing industry and the media are still grappling with issues of authenticity in the wake of the Mia Ballard affair. Having broken that story, the New York Times ran a piece describing “panic and paranoia,” and highlighting the potential inaccuracy of AI detection software, though without reckoning with the role of similar tools in its original reporting.
The Financial Times launched its 2026 Business Book of the Year award with an interesting round-up of how business book authors and publishers are evolving their products, or as the piece calls it, “the weird, AI-fuelled future of the book “: well worth a read, as there are useful signals about products, value propositions and reader relationships that are applicable outside business publishing.
Google finally caught up with its major competitors with the release of a macOS desktop app for Gemini—potentially useful for the many publishers running Apple hardware, though unlike the Claude and ChatGPT desktop apps, it is chat only and has no ability to run agentic workflows locally.
New data from Chartbeat show that declines in web traffic since the introduction of AI search have been sharpest for small publishers, which saw an average 60% decline in traffic for the last two years, versus 22% for larger publishers. Analysis in Axios suggests that for smaller publishers, focusing on brand-building and D2C relationships is now a stronger approach than traditional search optimisation.
This was originally published in my email newsletter. To receive weekly updates on how AI is affecting the publishing industry, sign up here.