Context Window 63
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 spoke at the Independent Publishers Guild Spring Conference this week (my notes from the event are here), discussing the AI landscape and in particular the impact on careers. In that presentation, I mentioned a new piece by Mark Ritson, one of the most interesting and opinionated writers on modern marketing. This week, he wrote about the collapse of the marketing job market. It’s a sobering piece that offers a clear but blunt picture for candidates, including:
And yet. This week I looked through nearly three dozen open marketing positions listed on The Bookseller and Publishers Weekly, including senior digital and online roles. Only two mentioned AI skills (props to Ingram and PRH). Maybe other publishing recruiters think it’s such a prerequisite that it goes without saying, or that publishing is somehow immune to general trends—but frankly, though I don’t wish for pressure on marketing jobs, I doubt it. Whether I look at this from the perspective of a publisher hiring the best talent, or an individual looking for career optionality in the wider marketing sector, I suspect there is a growing gulf between Ritson’s diagnosis and publishing’s current reality.
The most-shared piece in my feeds this week was an HBR article arguing that AI doesn’t improve task efficiency—it increases work intensity. Again, interesting, but worth pointing out that it was based on a study of a single organisation, which limits how far the conclusions can be applied to other organisations. I certainly haven’t seen the dynamics described in the piece in any publisher of comparable size that I’ve worked with.
Maybe you can help prove me right or wrong on those opinions: Helen King is running a survey on AI use in publishing for her own, excellent PubTech Radar newsletter. I’ll be very interested to see what usage patterns it highlights.
One of the other highlights of the IPG Conference was BBC culture and media editor Katie Razzall talking about the changing media landscape. On that subject, Digiday has an interesting piece on media companies optimising for formats that AI can’t easily scrape and compete with, such as live blogs and video. For book and journal publishers—by definition, slower media—it’s hard to see a direct parallel.
Another takeaway from the conference was news organisations using AI tools to analyse large volumes of newsworthy documents. Tax and transparency campaigner Dan Neidle has released code and a tutorial for the agentic AI search tool he is using. The approach would be easily replicable. If this kind of agentic search can navigate complex, badly-redacted public disclosures, imagine what it could uncover inside a publisher’s structured editorial or research archive.
ElevenLabs launched its Audiobooks product, part of its ElevenCreative suite of tools. It allows creation of audio using stock AI voices or a clone of an author or narrator’s voice, and offers detailed editorial controls and professional services for publishers with larger catalogues. I’m told that over 75,000 audiobooks have already been produced through ElevenLabs. It’s tied to a distribution service that reaches Spotify, InAudio and other sales platforms—but crucially not Audible, which will be a key commercial factor for many publishers, even if they have large backlist availability gaps.
In the same week, the company announced a new $500 million funding round, which values ElevenLabs at $11 billion and makes it, I think, the first UK-headquartered AI and media decacorn.
For an alternative lens on the same subject, a new working paper from the National Bureau of Economic Research suggests that the number of new books published increased by a factor of three between 2022-25, the same period as the initial diffusion of generative AI, but that among the top books in a given category, quality (as measured by user ratings) increased too.
Last week I wrote about Microsoft’s new Publisher Content Marketplace: this week, Amazon joined the party with plans for its own data exchange. This poses some interesting issues. Will Amazon’s existing publisher relationships make this better suited to book and journal publishers than Microsoft’s online and news-focused efforts? And building a successful multi-sided market with multiple buyers and sellers generally requires subsidising one side of the market to prime the pump—so early traction may not prove durable.
The UK government is obliged by law to publish an economic impact assessment of the options set out in its consultation on AI and copyright. UK readers in the creative industries have until 22 February to respond to a survey being carried out by Alma Economics on behalf of the Department of Culture, Media and Sport to inform this assessment. Please make sure publishers’ voices are heard.
Speaking of economic impacts, with many creative people worried about the effect of AI on their livelihoods, it is interesting to note that on the other side of the Irish Sea, the government is introducing a basic income scheme for the arts, after a cost-benefit analysis of a pilot indicated that it generated more than it cost.
Finally, more publishing than AI, but Bristol University Press & Polity Press (where I serve on the business advisory board) is advertising for a successor to CEO Alison Shaw, who founded the publisher in 1995. It’s a great opportunity for a scholarly publisher to shape a leading university press and work with a fantastic team.
This was originally published in my email newsletter. To receive weekly updates on how AI is affecting the publishing industry, sign up here.