And, of course, no similar call to suspend or boycott Iran, Russia, or North Korea in The Lancet. Fucking anti-Semites! (The World Medical Association disapproves the petition.)
If AI were treated like an ordinary consumer product, cost-benefits would have to be weighed, safety measures established, terms of liability determined. That's not what big companies want in general, esp. not re: groundbreaking tech whose long-term impact can't properly be assessed yet.
To lend the technology an air of inevitability and of inaccessibility to control makes the public and regulators doubt, thus, waste time, in which the AI companies can build infrastructures, establish pervasiveness in consumers' life, and eschew liabilities in absence of legal standards.
People love to talk about the "Tech Barons", "neo-feudalist" actors driven by libertarian ideology (cf. the "libertarian manifesto" by Luke Kemp), aiming to create autonomous city states without government regulations interfering in economic "liberty" (Peter Thiel).
Yes, these things exist. As does the transhumanist movement of human-machine merger and augmentation (Larry Page, Jeff Bezos, Elon Musk). At least as ideas. But these are irrelevant as well.
One might argue that having AI being exempted from the status of an ordinary consumer product plays into the hand of these libertarian and transhumanist aspirations. But for that AI is too groundbreaking and, because of that, still too vague to serve as a tool to bring about these fantasies.
And I think these people know that. So they stick to what they know best: Interrupt markets, scale quickly, define and dominate customer dependencies, try to evade liability or shift the burden on the public. Along their customary approach to lay claim on new territory while insisting to act without regulation, to suffocate any discussion on AI being a consumer product in need of regulation, and instead hyping up what a cultural and technological breakthrough that is, is sensible company strategy.
I'm aware that seen as an argument this is question begging: What has been the indicator of a thesis later turns out to become its evidence. So, strictly speaking, I haven't argued in favour of AI companies pursuing this strategy. But still, as an insight, it makes a lot of sense to me.
Thanks to Cal Newport for triggering these thoughts.
As long as anti-technology sentiments were still a cultural option, the fear of climate change peeked. Now with industries, military, and governments aggressively reinvigorating the cult of technology again, AI becomes the final frontier of horror. Not to forget that the last end time fantasy prior to climate change was mutual nuclear destruction in the Cold War, esp. in the early 1980s till the end of the USSR.
All three end time fantasies are based on the most horrific extrapolation of a pervasive cultural and technological trend of its time. This fantasy is not localized but always of planetary scale. All three are man-made inventions, having become uncontrollable and run wild.
Another characteristic of these end times is the occurrence of proselytizing movements calling people to repent before the Day of Judgement. And, obviously, it is THIS time that things are REALLY at stake. And if we not repent, then we face a global meltdown.
I don't deny dangers. Of course the world was several times close to nuclear exchange (the last time in 1983) and we managed to come through. Of course there is climate change (and the worst case scenarios RCP8.5 and its successor SSP5-8.5 have been scrapped). Of curse there is danger and risk with #AI. And I don't say that we will enclose AI like we did with mutual nuclear destruction or, in part, with climate change.
But what I do say is that undetected semi-religious urges keep nudging us to fall back into a script of the end of the world that has originally been written about 2,000 years ago. And that is not in itself good advice for risk assessment and risk management.
There may be good reasons why the seeming need for God-like punishment for unrepentant people return in secular times in various secular guises. Or to put it in the language of another context: There may be good reasons why in times of an alleged "Anthropocene" the wild and the wilderness return inside the human created environment called "the world" in form of unpredictability, chaos, neuroses and mental ailments.
But I find it important to highlight that our reactions to factual risks as to neurotic fantasizing always seem to be to return to the same old patterns and recipes already suggested in the Christian bible. Not that this were in itself automatically the wrong approach. But it is an approach by which we primarily enact a drama that probably caters to psychological needs and only secondarily something that may guide us to tackle the issues at hand.
A good example of this unquestioned attitude is the constant and trite humanisation of AI. When main researchers willy-nilly talk as if AI were (or could be) "conscious", have "intentions", "wants" and "needs", and, of course, "urges to destroy us", that already shows how people don't reflect on how their use of language posits qualities to world phenomena for which they have no evidence other than their unexamined projection.
There are indeed good arguments that humans perhaps cannot but project such qualities onto any and every agent that displays sufficiently complex response behaviour. (see e.g. https://bsky.app/profile/simsa0.bsky.social/post/3mhtefafdzk24 et passim). And perhaps machines may one day indeed be conscious (with or without a sense of self) independent of whether the underlying molecular structure is biological or cast iron. But in all cases this is playing with our fantasies and treating stuff so that it exhibits these fantasies.
"Everyone is an artist", Joseph Beuys once famously quipped (1969). And that lovely sentiment is quite different from Andy Warhol's reactionary claim that "Everyone will be famous for 15 minutes" (1967). Because what Beuys acknowledged as an essential part of the human condition, Warhol could only fathom and appreciate as a form of making one's usability and exploitability the most precious resource in a capitalist market economy.
What AI developers don't get is that they live the "Warhol Nightmare", not the "Beuys Dream".
They therefore misunderstand what they are dealing with and placate as nightmare what would far better and far more humanely be identified as another example of Man's age-old fascination with automatons.
Man always tried to understand himself with the help of metaphors derived from the most perfect machines he invented at a time. And always did he look half-scared onto the Golem he had apparently created, only to learn that, again, even this time he missed to create one.
Some, not all, discovery and recognition takes place when we lift the veil of metaphor. Esp. when it comes to technology and the changes we impose on environments, on infrastructure, on economic systems, it is recommended to know when and when we're not dealing under the spell of them. Because if we don't, we fall back to mutual fearmongering and scapegoating (and into the Warhol Nightmare). Which is never a good option when dealing with living beings and the planet as a whole. It's less fun, too.
. https://gnusocial.jp/attachment/468c884019248ef2d872716ef6b7bcdc1b80fd16f0e46740670b8c1c0fa686c5/view
"[Pakistani PM Sharif] said both sides have declared and [sic] 'immediate and permanent termination of military operations on all fronts, including Lebanon.'"