The cop-out answer is, ‘Yes there is, but we mostly don’t know yet what it will be.’ So this post is more about the qualities of books and articles that others are writing, but with a political spin.

For me, and many others, AI will change society as profoundly as did the steam engine and the electricity grid. This anticipation has led to boosterism and to a bandwagon of snake oil applications. A few of the more thoughtful commentators are reminding us of the old adage, ‘when the gold rush starts, sell shovels.’

Currently I am ploughing through a number of the better AI textbooks. And there is are many books out there which are, in my view, no more than paperback versions of clickbait. AI and Home Decorating – I’ve just made that up, but you get the idea. Currently I trawl the bookshops near to universities in London, where some have created new shelved sections for AI books, most of which can be readily skipped over.

AI books are also out of date before they are printed, but that said, for me the best in class at the moment is ‘Mastering AI’ by Jeremy Khan (2024), written by a tech journalist who knows (a) tech and (b) how to write about tech for the politically smart generalist.

Because my bachelors degree was in Computation I’m also reading a translation of a German textbook, which is good, but not for a general audience.

I would suggest that, if you want to follow AI developments you need to read political economists as well – the dark doings of tech bros in Silicon Valley and Seattle will shape it as much as the raw science will.

I also humbly suggest that people with a humanities background are at a disadvantage currently in terms of AI. My masters degree was in the humanities, so I hope this is taken in the spirit it was intended. I can imagine some universities where the humanities profs are glaring out of their windows at the sciences block, and committees are busy fixing last year’s problem.

I guess a wider issue is, how best can analogue staff teach a born-digital generation. For example, the Elicit website uses AI to help postgraduate researchers do what I would still call a literature review, and to share their latest papers with other researchers. And embedded AI will probably reach a point soon when it is no longer worth commenting on. For example, we don’t say that cars or washing machines or hair dryers have embedded electricity.

On the social and environmental impacts of AI companies – such as demand for power and cooling, trawling of intellectual property without permission, job displacements – I think most people argue backwards from whether they like (or worship) tech or not.

The pro side are saying, everything will be fixed and these are just teething problems.

The anti side are saying, it will add to inequality and exclusion, another aspect of creative work in society being enclosed and strip mined for profit by the tech bros. For example, you can imagine that somewhere in a garage in Hollywood is a geek or two working on an app that turns a 200-word prompt into 90 minutes of film, ready for viewing.

Jeremy Khan covers these pro and anti points in detail and without boosterism, in my view. His conclusion, which I paraphrase here, is that the future path of AI will continue to throw up some surprises, but how we deal with all this is a political choice where governments around the world are struggling to regulate tech companies, and tech companies are always trying to ‘innovate’ to evade laws and taxes written for an earlier age.

You might call it unregulated capitalism.