Well that is a new job title.

Now obviously I meant ‘What’ and ‘speaks’.

Thanks Matthias (also for your original comments)!

Thanks Abby! Sharing offices is sometimes a good thing.

Link again, for good measure: journals.sagepub.com/doi/full/10.... It's fully Open Access as well, so anyone can read it.

This was a really fun paper to write - we spent several hours just chatting in the office Cristobal and I shared at the LSE (and the remaining hours at the pub), and came up with a paper that none of us could have written individually. We'll have to go back to the pub for the sequel.

TL;DR? AI may wall be a transformative technology, but we should be wary of (many of) the claims that are being made about it. These claims are both political and depoliticizing, and risk skewing the debate about what kind of future we should pursue.

The appeal to urgency further reduces the space to imagine, let alone bring about, alternative futures. Democracy needs time and collective involvement, while emergencies concentrate power in the hands of experts and decision-makers.

Many of these claims overstate the speed at which AI is developing & create narratives of certainty where scientific knowledge is significantly less certain. They are, in other words, not technical claims, but political ones that colonize the future & seek to spur action in their desired direction.

Finally, we look at temporal themes. Discourses on AI portray it as a force in motion, which is changing societies and will do so even more quickly and radically in the future. The claim of a sudden acceleration prepares the ground for demands to act urgently to speed it along, or halt it.

These assumptions, we argue, are problematic to the extent that they uncritically reflect key pillars of a certain form of liberal capitalism and its depoliticizing tendencies.

A third set of themes we identify are economic ones. Economic discourses on AI have tended to portray the development of the technology as essentially linked to a specific conception of what economic reality is and, perhaps more importantly, to a specific conception of what that reality ought to be.

Presenting AI in such terms ascribes it agential qualities that remove responsibility from humans for creating the technology in the first place, and presents human agency as purely reactive. Doing so, we argue, it promotes a form of politics that leaves very limited space for political choice.

We then discuss agency. Here we find a peculiar situation of agency without responsibility. One the one hand, AI is described as potentially escaping human control. At the same time, the development of AI also becomes an opportunity to reassert human agency and use it to direct the future of AI.

At the same time, those who say that AI affects humanity underplay the differential effect of AI on individuals and collectives. This makes it harder for specific groups to mobilise around the aspect of AI that affects them.

We start with claims about the nature & comparability of human and artificial intelligence. We argue that claims about the comparability of human and artificial intelligence fail to recognize fundamental differences between human and artificial intelligence.

We develop this argument looking at four sets of themes in contemporary public pronouncements on AI, and their underlying assumptions.