Essay by Eric Worrall
Soros leading the pivot from the fake climate crisis to the fake AI crisis.
AI is at the root of the world’s ‘polycrisis’
Climate change and the war on Ukraine are also threatening democracy, according to billionaire and philanthropist George Soros.
George Soros
We are living in troubled times. Too much is happening too fast. People are confused. Columbia University economic historian Adam Tooze has, indeed, popularised a word for it. He calls it a “polycrisis”.
The polycrisis has many sources. In my opinion the main source of the polycrisis afflicting the world today is artificial intelligence. Climate change comes second, and the Russian invasion of Ukraine qualifies as the third. The list is much longer but I’ll focus on these three. That should help reduce the confusion.
Artificial Intelligence
…
Shortly thereafter, Geoffrey Hinton, who is generally considered the godfather of AI, resigned from Google so he could speak openly about the risks posed by the new technology. Reversing his previous position, he took a very dim view of AI. He said it could destroy our civilisation.
…
What Hinton said made a big impression on me. Indeed, AI reminded me of Goethe’s poem The Sorcerer’s Apprentice. The apprentice is studying magic but doesn’t fully understand what the master is teaching him. When the master orders him to sweep the floor, he applies the magic words to a broom. The broom obeys him, but the apprentice can’t stop the broom from fetching buckets of water to sweep the floor and the house gets flooded.
…
Read more (paywalled): https://www.afr.com/technology/ai-is-at-the-root-of-the-world-s-polycrisis-20230612-p5dfui
I predicted this pivot back in 2017. The political utility of the fake climate crisis is all but spent, climate crisis rhetoric these days mostly only works on lefties who already planned to vote socialist. But AI scare stories have bipartisan appeal. And we’ve already been well primed for the fake AI crisis, by science fiction / horror stories like “The Terminator“.
I’m a software developer of over 30 years experience, who has personally written AIs. I’m not a “father of AI”, but I know a thing or two. I disagree with the AI apocalypse narrative.
The Sorcerer’s Apprentice is a powerful metaphor, but in my opinion it is the wrong metaphor for rise of AI. A better metaphor is the arms race between computer virus creators and anti-virus companies.
Every so often a virus writer manages to slip in some blows, but our computers are mostly safe. Install the anti-virus, keep the subscription and software patches up to date, and you can go about your daily business.
In a similar way, whatever capabilities AI gives to miscreants, or whatever aberrant capabilities AI develops on its own, it will be treated like yet another computer virus, to be countered by AIs primed to monitor and respond to threats.
Is it possible someone could leap ahead and develop unique capabilities, or that an AI could develop a novel form of attack? Of course it is – look how Microsoft’s ChatGPT integration caught Google flat footed, an example Soros mentioned elsewhere in his article.
But who thinks this setback will be the ruin of Google, and all the other AI companies? By the end of this year the world will be awash with ChatGPT lookalikes, created by tech companies spending 10s of billions, whatever it takes, to stay in the race.
AI will create tremendous changes to society, some marvellous and others which many of us will find deeply disturbing. On the positive side, there will be glorious advances in medical science. Advances like medical immortality and cures for currently intractable diseases are almost within reach.
But AI will also bring challenging disruptions. Among other things, I foresee a future AI version of today’s schoolkid “trans” conflict, with progressive politicians demanding kids should be allowed to augment their brains and bodies with AI implants without parental consent, to avoid the trauma of feeling inferior to their augmented classmates.
But we’ll get through all that and more. Many of us alive today may live to see an age of marvels – an age of fulfilment and joy which today’s world can only barely glimpse.
In the meantime, we need to stand firm against this pivot to AI scare stories, just as we stood against climate scare stories and Covid lockdown scares. Because at the base of the AI fear campaign will be the very same people who are currently at the base of the climate crisis movement: meddlesome fools who believe they know how to run our lives better than we do.
My only “concern” about “AI” is that it makes it so much easier for propagandists to get their messages out to the public, in a conversational and convincing way.
And too many will believe the horse shit that AI “bots” trained on “approved” data bases will spew as if it is unbiased and authoritative, when it is anything but.
A classic recent headline: “Artificial Intelligence Cites Artificial Case Law.”
There is a law firm in serious hot water over that fiasco.
Never mind existential threat, AI is an existential opportunity. AI is perfect as a replacement for “climate change” as a focus of hysteria. Fight against the doomsayers.
https://maxmore.substack.com/p/against-ai-doomerism-for-ai-progress
There is a simple demonstration that “AI” has gone down the wrong track. Show kids a horse, one horse, and, of course, tell them it’s a horse. Then show them a donkey, and they’ll ask “Is that a small horse?” Then show them a zebra and they’ll ask “Is that a striped horse?”
People do not need to be trained with millions of examples.
Actually we do – a human who never saw a zebra before might ask the same question. The millions of examples we receive via interacting with parents and teachers.
Where humans have the edge, and its a big one, is we have instincts honed by a billion years of evolution. AIs do not have these instincts, which makes them act really dumb in ways humans would never consider.
Here I disagree with you. Every tool is two edged. Call them AI or call them something else, thests types of software put great power into the hands of the surveillance state, to say nothing of the corporate overlords. While they may have all sorts of uses in many areas, they will be used against the population at large and most especially against individuals who don’t accept the reigning groupthink. There is no need for the programs to have a mind of their own.
AI is already ubiquitous, your vehicle satnav contains an AI, as does your smartphone, with predictive text and all its other features.
You are absolutely right, the surveillance state will be enhanced by AI. But mostly they will be distracted by other surveillance states, 99% of their energy will be taken preventing foreign interference. The remaining 1%, people are smart, we will figure out ways to deflect government intrusions into our lives.
No doubt just like the majority of China’s peoples are doing.
The Chinese Great Firewall has turned out to be quite porous.
Far more dangerous than AI are (in no particular order):
And because they are dumb, in the true sense of the word, they do not know when they are making a mistake. This could be truly disastrous when used in our military systems. One doesn’t need to paint the horrendous scenarios that may occur when dumb Defence personnel comes to rely on dumb computers to run their responses to perceived threats for them.
A couple of days ago I decided to test ChatGPT. I expected it to have been trained with the mainstream “consensus” pseudo-science and all the media blurb flowing from it, which it would encounter on the internet. I was curious to how it would respond to challenges. Maybe it would point out flaws in some of my ideas or bouncing it back and forth may strengthen my arguments.
a) I first asked in ENSO could have an effect on long term climate.
It said yes, but then produce a paragraph of the usual waffle about water “sloshing” which did not have any bearing on whether there was a long term effect or not.
b) I asked how La Nina affected OHC.
It said colder SST caused more heat loss to the atmosphere and thus reduced OHC. So I asked it to explain how cooler SST caused more heat loss, not less. It appologised for the “confusion” and said that was wrong, it was in fact due to increased winds and increased upwelling of deep ocean waters mixing with surface waters thus increasing OHC.
I pointed out that deeper water was colder and OHC could not be increased by simple mixing only by heat exchange with another medium.
It was again sorry for the “confusion” ….
After three stupid mistakes which were categorically wrong, not even controversial, I asked that after 3 mistakes I could have any confidence in the responses it was giving me.
I again apologized pointing out that it was a linguist model and not “infallible”.
What is most scary about all this is not capabilities of AI but the fact that people are prepared to believe anything produced by a computer is objective and beyond question.
Ask AI about climate science and it will explain to you it has six fingers on each hand.
‘glorious advances in medical science’???
Ha ha ha!
I agree with the writer’s assessment of AI.
I also am a software developer with over 30 years. I haven’t been contracted to write AI software, but I’ve researched it enough. Never mind the question of what is thinking or consciousness. Scary AI stories are based on the idea that the machine could develop an ulterior agenda, a self interest of its own. Anyone believing that doesn’t understand that a machine will always be just a machine. The malevolent entity is the human being programming the machine.
But there is a legitimate concern for the appropriate use of computers and software. Do you want a computer deciding your case in court? Do you want a computer deciding who will be your elected official? Do you want a computer driving your car or piloting your airplane? I think not.
The scare about AI will lead to government oversight of computer programming. Software engineers will eventually have government agencies looking over their shoulders.
Computers have been flying planes for decades.
The newest ones can both take off and land planes.
There is some potential for AI to propose medical diagnoses from diagnostic (Lab, imaging, etc) results after being trained on millions of patient charts. The weakness in this scheme is a part of the chart known as the “problem list”, which lists current and past diagnosed problems without metadata (currently under treatment?, onset and duration of problem? etc.). Problem lists are also incomplete — I’m 70, but it could be relevant that I had mononucleosis as an 18-year old.
This scheme would require a universal patient identifier, something many civil libertarians fear. AI complicates this further — if someone suffered childhood sexual abuse, might not an AI disqualify them from employment that involves access to children?
The danger in AI depends on what actions we allow them to take in real time. I’m not opposed to AI synchronizing traffic lights or managing air traffic control. An AI that involves value judgements, however (like a Chinese-style social credit scheme) is scary, though.
One difference I’ve noticed between veteran programmers and fresh out of school ones, is that when interviewing the client, the veterans have learned how to ask questions in order find out what it is that the client really wants.
I certainly agree that the climate crisis is a scheme to give the elite more control over the rest of us. But I think AI really has the potential to get away from us. The things the computers can do they do much better than we do, i.e. doing time consuming calculations in a fraction of a second or playing chess. And certainly there is a good chance AI will bring great benefits like finding cures for incurable diseases. Some folks have referred to AI as having godlike intelligence. Would we be able to control a god? I doubt it.
There will never be medical immortality, unless it means something very limited and nothing at all to do with immortality.
How many people in our Western societies are paid to be confident even if clueless and often wrong?
AI can express confidence (and be far cheaper).
Talking heads will be needed for TV… for now.