I have seen the direct impact of AI on my kid's career choices, dropping courses they see no future in. A kill switch is critical. Humans must retain ultimate control. The movie Fail Safe was made in 1964 and remains true today. Otherwise the day will come when WOPR asks: "Shall we play a game?"
I find it terrifying, far beyond fretful. For all the good it could do, the worst it could do is catastrophically disastrous for the planet and everybody on it. Heaven forbid any form of oversight should be allowed, just in case a few incredibly rich people have to act ethically.
Top article. Humanity is face with the challenges of irreversible climate change and the possibility that it loses control of artificial intelligence. Are we heading towards extinction pushed by these powerful forces?
Amazing NB, this is a deep thought provoking article you have written. There is so much to keep up with these days but thankfully you are there to navigate the ''What about you?' boomer cohort' of which I'm a member through uncertain (hasn't the world always been that way?) times. Me, personally, I've always adhered to a principle of 'fairness' to/for all, whatever that may mean?
Too easy to confuse “unstoppable” with “ungovernable”. Yes it’s likely that AI will persist for some time, depending on which comes first, some new technology that renders it superfluous or societal collapse.
For every technology we’ve invented we have forgotten more.
It’s also easy to forget that there are three main players in this particular phenomenon, owner builders, governments and users.
The standard method for owner builders to divest themselves of responsibility for bad outcomes is to claim 1. That the state of affairs they want is inevitable. (Maggie Thatcher’s “there is no alternative”.) 2. That if things go wrong it’s the fault of the users. (Any lobbying advertising campaign from cigarettes to gambling to not disposing of plastic waste properly. ) 3. That any attempt to stop or regulate whatever they are selling will bring the economy to its knees.
Governments, as we are prone to forgetting, can regulate. They regulate to produce the market rules that allow markets to operate. They can if they wish and are not too scared by lobbyists powers to whip up a furore, actually ban some products and activities.
Users usually don’t have much say. They are the ones who are the source of the data that makes AI possible and the ultimate source of the money that floods upwards, but almost never trickles down.
Really we have to stop thinking that everything that AI does is somehow down to the technology itself and not down to the people who stand to benefit.
The well of AI-fandom is shallow in my neck of the woods. Across generations concern is deep. The US is a poor place to look to for leadership on this. For all its faults, the EU is the most likely jurisdiction to impose meaningful limits on AI - and is big enough to make them count.
Total reliance on automated kill switches creates a dangerous illusion of control that masks the underlying loss of human agency. We built systems that operate beyond our immediate oversight, yet we still expect to maintain final responsibility for the outcomes. The next phase requires boards to formalize the return of decision authority to human operators. By codifying these boundaries, we stop treating autonomy as a liability and start managing it as a core operational asset. We must move past the passive acceptance of technical defaults to actively reclaim the governance of our own infrastructure. This shift turns the current anxiety into a clear mandate for structural reform at the highest levels of the organization.
Readers of Science Fiction will be familiar with the Laws Of Robotics written by Isaac Azimov decades ago. A prophetic voice. Certainly, limitations are necessary. After all, we make laws limiting our own powers and behaviours, why does anyone think that artificial intelligence and algorithms ought to be exempt? We are already seeing social and psychological harms imposed by social algorithms media and belatedly beginning to rein them in.
It’s time that our governments took proper responsibility for governing AI and algorithms. No entity or person ought to be above the law.
Yuval Harari is correct in the concerns he expresses regarding algorithms and AI versus Humanity.
Looks like another book in gestation. If so, better get it off to the publisher before another global change/catastrophe sends the West roiling again. No end-points foreseeable, certainly not directed by whatever is left of universal social democracy.
I have seen the direct impact of AI on my kid's career choices, dropping courses they see no future in. A kill switch is critical. Humans must retain ultimate control. The movie Fail Safe was made in 1964 and remains true today. Otherwise the day will come when WOPR asks: "Shall we play a game?"
Highly recommend this interview by Rory Stewart of Matt Clifford and Jack Clark (a co-founder of Anthropic). My takeaways are that AI is unstoppable, that it is not too late to control AI, and that if we do, the good outweighs the bad. https://open.spotify.com/episode/3O7eWIrWY1kWIVarXxoLtj?si=7LAa58saTbKFvp90Gzdr_Q&utm_source=copy-link
Chilling
I find it terrifying, far beyond fretful. For all the good it could do, the worst it could do is catastrophically disastrous for the planet and everybody on it. Heaven forbid any form of oversight should be allowed, just in case a few incredibly rich people have to act ethically.
Top article. Humanity is face with the challenges of irreversible climate change and the possibility that it loses control of artificial intelligence. Are we heading towards extinction pushed by these powerful forces?
Amazing NB, this is a deep thought provoking article you have written. There is so much to keep up with these days but thankfully you are there to navigate the ''What about you?' boomer cohort' of which I'm a member through uncertain (hasn't the world always been that way?) times. Me, personally, I've always adhered to a principle of 'fairness' to/for all, whatever that may mean?
Too easy to confuse “unstoppable” with “ungovernable”. Yes it’s likely that AI will persist for some time, depending on which comes first, some new technology that renders it superfluous or societal collapse.
For every technology we’ve invented we have forgotten more.
It’s also easy to forget that there are three main players in this particular phenomenon, owner builders, governments and users.
The standard method for owner builders to divest themselves of responsibility for bad outcomes is to claim 1. That the state of affairs they want is inevitable. (Maggie Thatcher’s “there is no alternative”.) 2. That if things go wrong it’s the fault of the users. (Any lobbying advertising campaign from cigarettes to gambling to not disposing of plastic waste properly. ) 3. That any attempt to stop or regulate whatever they are selling will bring the economy to its knees.
Governments, as we are prone to forgetting, can regulate. They regulate to produce the market rules that allow markets to operate. They can if they wish and are not too scared by lobbyists powers to whip up a furore, actually ban some products and activities.
Users usually don’t have much say. They are the ones who are the source of the data that makes AI possible and the ultimate source of the money that floods upwards, but almost never trickles down.
Really we have to stop thinking that everything that AI does is somehow down to the technology itself and not down to the people who stand to benefit.
The well of AI-fandom is shallow in my neck of the woods. Across generations concern is deep. The US is a poor place to look to for leadership on this. For all its faults, the EU is the most likely jurisdiction to impose meaningful limits on AI - and is big enough to make them count.
Total reliance on automated kill switches creates a dangerous illusion of control that masks the underlying loss of human agency. We built systems that operate beyond our immediate oversight, yet we still expect to maintain final responsibility for the outcomes. The next phase requires boards to formalize the return of decision authority to human operators. By codifying these boundaries, we stop treating autonomy as a liability and start managing it as a core operational asset. We must move past the passive acceptance of technical defaults to actively reclaim the governance of our own infrastructure. This shift turns the current anxiety into a clear mandate for structural reform at the highest levels of the organization.
https://cyrilsimonnet.substack.com/p/we-signed-away-our-autonomy-now-the
Readers of Science Fiction will be familiar with the Laws Of Robotics written by Isaac Azimov decades ago. A prophetic voice. Certainly, limitations are necessary. After all, we make laws limiting our own powers and behaviours, why does anyone think that artificial intelligence and algorithms ought to be exempt? We are already seeing social and psychological harms imposed by social algorithms media and belatedly beginning to rein them in.
It’s time that our governments took proper responsibility for governing AI and algorithms. No entity or person ought to be above the law.
Yuval Harari is correct in the concerns he expresses regarding algorithms and AI versus Humanity.
Looks like another book in gestation. If so, better get it off to the publisher before another global change/catastrophe sends the West roiling again. No end-points foreseeable, certainly not directed by whatever is left of universal social democracy.
John Quiggin's https://substack.com/home/post/p-208422685 worth a look, also Branko Milanovic's https://branko2f7.substack.com/p/sketching-the-new-dysto-utopian-world.
(We are glued to 'Person of Interest' on Netflix, set in early 2010's and startlingly prescient of Alex Karp and the other tech overlords.)