One convenient way of keeping a weather eye on developments in Artifical Intelligence is to follow the feed from Marshall Kirkpatrick’s AI Time to Impact. This daily stream of alerts sends its followers brief editorial summaries and links to hot news across the whole field of AI, in business and enterprise, academia, government and regulation, chips and code. Usually, Marshall will add some estimate of the time-reference of the breakthrough, story, milestone, or research, in question. A measure of the timeliness or ‘time to impact’ of what we are seeing. These measures are: short, medium, occasionally long or, even, ‘TBD’ (to be determined).
The overall global impact of the AI boom and investment cyclone that we have unfolding in front of us is quite clearly in the TBD category. And this must be true for the profound impact on specific industries and cultural institutions : Hollywood, the music industry, the visual arts, science and medicine, the worlds of code and language, newspapers and magazines included. Yet, it may be useful and timely to make some estimate or broad guesses as to the probable impact of the AI boom on the magazine industry. Here are three guesstimates:
(1) Time to impact: SHORT
Advertising within magazines is likely to be much less important than subscriptions now that AI is moving to grab consumer attention.
Corollary: digital magazines in an AI culture will be increasingly reliant on subscriptions or membership communities. Magazines have already caught wind of this change from the early warning that Google and Meta have given about the demise of third-party cookies. AI promises to be very big in advertising (partly because that is where Meta, Amazon and Google will be driving the technology) but it will not be targeting consumer audiences collected in digital magazine communities, and any new advertising margins in those communities will be sucked up by the oligopolies. Get those digital subscriptions moving!
(2) Time to impact: MEDIUM
Magazines will increase their value by being reflective and considered
One of the key features of the magazine medium comes from the periodical nature of periodicals (sometimes called serials by librarians). For their audience, and their editorial process and relevance, magazine content has a periodical rhythm: weekly, monthly, quarterly, bi-annual or even unpredictable issues (usually a mistake). This predictability is sometimes seen to be disadvantageous, and the disadvantage may become a good reason for a magazine, of a certain sort, turning itself into a continuously updated website or database (reference to the clear commercial success of Auto Trader, Antiques Trade Gazette, and Which? magazine: but notice that many troubled titles have walked this route and relatively few have succeeded). The periodical, serial, deadlined, edited, usually fact-checked, and issue-based, pattern of content flow in magazines and newspapers is deeply advantageous for the medium: as we noted above this broadly editorial and design work (even the mastheads, correspondence columns, the print advertisements and by-lines) is in effect creating reliable metadata within the published content. Furthermore, when there are almost instantaneous forms of free news-flow (X/Twitter, TikTok, CNN, Facebook News: if that will still be there when we need it?), the advantages of edited and reflective comment, design, illustration, reportage and analysis can make for a premium product. The episodic, temporal, structure is as valid and valuable for fashion magazines or poetry reviews as for obviously analytic or controversial titles such as The Atlantic, the Spectator or Prospect. The recent wave of Large Language Models have found themselves with a host of legal problems through appearing to ingest and reuse copyright material without permission. But it was not just the quality and depth of New York Times journalistic content that enriched the training schemes of Chat GPT and its peers. Each story in a newspaper carries its own provenance and date. Newspapers are rich in valuable metadata. AI of the LLM type falls over when it has no sense of history or pertinence. Striking deals with reputable and time-reliable publications such as NYT, WSJ, and the Financial Times is essential also for the flow of future content. It will be required on a continuous and co-operative basis. When AI of that omnivorous and reliable LLM kind is working widely, and with trust, it will be because rights deals have been done. If AI becomes as omniscient and omnivorous as its backers suppose, deals will need to be done, with more specialist titles: The Economist, The Times Literary Supplement and with Opera, the New York Review of Books, and The Bookseller. In their admittedly narrower fields of expertise, those titles are as authoritative as the New York Times.
(3) Time to impact: LONG
Magazine publishers will realise that a thoroughly digital magazine operating in AI cultures needs to have a deep retrospective and prospective archive. Looking forward as it searches back.
Exact Editions has always operated on the principle that a digital magazine should have an archive, and so far as possible it should carry its archive with it, as it moves along. As the magazine inches forward, on its periodical progress, by week, by month, by quarter and accumulating annually, or even by decade, the databased content grows and the subscription content available to its readers deepens. Magazine digital access should include access to available (ie databased) back issues; perhaps Exact Editions made this choice because archives and back issues belong with magazines and periodicals as roots belong with trees. We might claim that this was an aesthetic choice (well it was that — we like back issues as much as any publisher or reader) but it was also a choice to some extent forced by efficiency. It is much easier for readers and suppliers, above all for the manager of access, to have one complete product with one class of audience or subscriber — to welcome everyone with a subscription to have access to all available issues.
If the SHORT and MEDIUM term predictions, already tabled, turn out to be correct (admittedly two big enough ifs!) it is conceivable that magazine publishers and magazine readers will begin to think of digital magazines in a new way. Publishers will be thinking of their readers as having access to a valuable and ongoing content flow (which perhaps may be simultaneously subscribed by AI services or training regimes) and this will be as much predicated on the future flow of content as on the accumulated, databased and curated back issues. Readers, subscribers or members, then become important and vital to the publication because they assess, value and appreciate knowledge, opinion, style and design that has been accumulating (yes, they search and browse back issues) but they also look for, pay for, and anticipate the content of future issues. The AI:Time to Impact service classifies as ‘LONG’ any development that will have its prime time after 2026. Making any AI dependent prediction that stretches into 2026 is a foolhardy business, but here goes.
For many years publishers who are making a serious pitch for print-based advertisers usually produce an annual Media Kit (some current examples here). Why have we not already invented the notion of a Subscriber Kit? Especially useful if there turns out to be a market for long term AI deals? If the search for long term supporters and subscribers accelerates we can expect to see publishers producing and committing to Subscriber Statements or Sustainer Principles. Subscriber Kits — though they will not be called that — should also take advantage of the fact that an easily navigated archive enriches prospective commitments (eg to the next season in golf, football, design, fashion or social features). Explicit commitments as to the range of content or editorial features that can be expected, perhaps a statement as to the role to which AI is. or is not, involved in the editorial/design/browsing process, and perhaps also the explicit acknowledgement of dollar funding premium subscribers, and sundry Artificial Intelligence backers. Expect also to see sustaining supporters recruited to pay for the accurate production of complete archives for hobbyist magazines. Expect also to see magazine publishers looking for lifetime digital subscribers. Exact Editions has for some years been offering perpetual and ongoing subscription access to some of the complete archives that offer institutional subscriptions. In an era of distributed AI and real-time search, a subscription for Permanent or perpetual Access can be both prospective and retrospective. And it will be the commitment to forthcoming content that makes a magazine most useful and most valuable to all its potential market, intelligences that are mostly human but by 2026 increasingly artificial.
Leave a Reply