The Terminator films showed how AI, as developed and used by the National Security State, could lead to a nuclear apocalpyse and a reverse 'Butlerian jihad' where machines wage a war of extermination against humans.
Clearly, there were dangers in Governments developing and using AI.
Equally, a private corporation, greedy for profits might ignore the need for safeguards & unleash an AI which becomes self-aware and pursues its own goals. It might decide to disintermediate humanity- destroy it- so as to get more resources for itself.
Currently, we are aware that there are two ways to develop AI
1) the American way, where giant corporations are in the driver seat. The advantage here is that Governments may find regulation easier.
2) The Chinese way, where there is intense competition between a number of small players. Obviously, the Chinese state has direct control over its citizens in a manner not feasible in 'Rule of Law' polities.
How can an AI apocalyse be averted? The answer is the same as the question of how internet viruses can be contained. There has to be an 'arms race' (or Red Queen race) between AI bad hats & AI white hats.
Writing for Project Syndicate, Kaushik Basu takes a different view. We must kill our own golden goose so that China can become all powerful.
How to Stop the AI-pocalypse Sep 25, 2026 KAUSHIK BASU
AI could very well end democracy, if not humanity, yet efforts to slow its development may fail as companies and countries race to stay ahead.
What Basu means is 'free enterprise may destroy civilization. Let's destroy free enterprise. What's wrong with quietly starving to death while yielding more and more territory to China? After all, that was the outcome for Nehruvian India in the early Sixties. Back then, the Bengali buddhijivi mathematical economist had power and influence. Once a country's material standard of living starts to rise, it tells Bengali intellectuals to fuck off. Only Punjabis or Gujeratis are worth entrusting with policy making power.
Preventing such a future will require international coordination
which is impossible
and taxing the extreme wealth that could put unprecedented power in the hands of a tiny elite.
Nothing wrong with taxing factors in inelastic supply. But the wealth created by AI is highly mobile between jurisdictions. This means it is elastic & thus won't bear the burden of any tax.
ITHACA, NEW YORK—Over the past decade, scholars and journalists have been writing about what the march of digital technology means for humanity with a mixture of hope and anxiety: hope because of the rise in living standards
if total factor productivity rises- sure. But AI is also a consumer good. What is it complementary to? I asked an AI which replied ' supermodularity arises explicitly with respect to the consumption and usage of AI. In economics and game theory, supermodularity defines a relationship of strategic complementarity. It occurs when increasing one variable increases the marginal benefit of increasing another variable (the "more-gets-you-more" dynamic). [1, 2]
With respect to AI consumption, this phenomenon operates across several distinct dimensions:
1. Complementarity with User Knowledge & Capability
Recent economic and management research modeling the demand-side value of AI formalizes a strategic complementarity between computational resource consumption and consumer knowledge accumulation. [1]
The Dynamic: An uneducated or low-skilled user gets baseline utility from a generative AI tool. However, as the consumer invests time in learning prompt engineering, domain-specific logic, or workflow orchestration, their ability to extract value from the AI grows exponentially. [1, 2]
The problem is what AI substitutes for. The middle class professional or 'knowledge worker' may be wiped out. The Pareto distribution gets 'fatter tailed'. The 'Matthew effect' kicks in. The rich get less numerous but have greater command of resources.
that the digital revolution has made possible, and anxiety because it has deepened social divisions and destabilized politics around the world.
This is the notion that Putin's bot army has turned people racist by spreading 'fake news'.
But anxiety has recently taken an existential turn, as the industry’s leaders have begun warning that AI could very well end humanity. Most notably, Evan Hubinger, Anthropic’s alignment science lead, said there’s a greater than 10% chance that AI will kill all humans within the next decade.
He wants more resources for 'alignment'. As a 'safety' guy, he benefits if people think there is a catastrophic risk.
Legislatures are likely to want higher concentration in the sector because, they believe, it would make their own 'oversight' more effective.
Similar, though somewhat less alarmist, warnings have come from OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Bill Gates, and other prominent tech figures, all of whom have called for slowing down AI development.
So as to lock in their first mover advantage. The problem is China. They have risen far more rapidly than anyone in the West anticipated. Moreover, once they produce more and better quality of their own high end chips, the West has no threat point. Indeed, the shoe may be on the other foot.
Naturally, these warnings have fueled concerns about what AI can do to us rather than for us. AI models can already build more advanced AI with little human involvement, raising the prospect of “recursive self-improvement” that unleashes forces we have little ability to understand or control. The Hugging Face breach, in which hundreds of AI agents that were supposed to work independently found ways to communicate and started acting as a swarm, offered a disturbing preview.
Hugging Face was started by a couple of French dudes. Why didn't they do it in France? The answer, at least partly, is that France has high taxes.
All this brings to mind Bertrand Russell. In his famous critique of induction, Russell noted how chickens that get used to being patted every morning are in for a shock when, one day, the farmer—presumably having invited friends for lunch—wrings their necks instead. As he observed, “more refined views as to the uniformity of nature would have been useful to the chicken.”
How would it be useful to the chicken? The answer is, it wouldn't be useful at all. That is why there was no selection pressure for 'more refined views' amongst domesticated fowl. On the other hand, there are birds in New Guinea which feed on toxic food items so as to make themselves undesirable for human consumption.
I share Russell’s skepticism of induction but am not persuaded by his advice for the chickens. If the chickens followed his advice, they would simply live their last days in helpless panic. It is different for humans, since, unlike Russell’s chickens, we have the power to modify our own behavior.
We are a self-domesticated species. This is because solving collective action problems contributes to inclusive fitness.
And there is no shortage of proposals for how we should curb the excesses of AI, from pausing the development of frontier models to banning superintelligent AI altogether.
Why not just surrender to China? When Basu was young, smart kids in Calcutta were saying 'China's chairman is our chairman'.
As sensible as these proposals may be, they may not work. Game theory, and the prisoner’s dilemma in particular, tells us why.
No. Both are irrelevant. There may have been a time when the West thought it could stop China from moving up the value chain in IT & AI. China started to close the gap from 2019 onward. Now they aim to implement AI across 90 percent of the economy by the end of the decade.
If each corporation is left to act unilaterally and in its own interest,
they will form a cartel.
we may end up with the worst outcome for everyone.
Cartels artificially reduce output and innovation. Maybe that's a good thing.
What is needed, as Amodei has argued, is “industry-wide coordination.”
While China takes over. Sadly, this means the national industry disappears because China can do 'dumping' because it has more economies of scope and scale.
Yet even that may not be enough, since AI is also a geopolitical problem. If the United States exercises restraint, China could surge ahead.
This has already happened. I suppose the Taiwanese will accept re-unification as inevitable. This means China's chip making capacity rises yet further.
Some commentators have seized on this fear to argue that the US must not hold back on AI development. That would be a mistake. The US should work with China to reach a global agreement, but even if that doesn’t happen, there is no reason to charge ahead at full speed.
No agreement with China means the possibility of a military conflict or, if China is now too strong, then an economic war of embargos and sanctions. Sadly, China has established threat-points of its own.
If AI is potentially lethal, the side that does not 'charge ahead at full speed' is the one without a threat point. This means the 'Nash bargaining solution' is for it to bend over and take it up the ass.
It is worth remembering that the only time atomic bombs have been used in war was when the US had a nuclear monopoly.
Because it 'surged ahead'. Then the Soviets caught up. Basu, cretin that he is, is arguing for 'surging ahead' at all costs.
It sounds Machiavellian, but nuclear restraint took hold only after the Soviet Union, and eventually other countries, acquired the bomb.
No. Japan had already been defeated. Nukes just gave them a way to surrender without too much loss of face.
The same logic may apply to AI.
It does. That's why there is no substitute for 'surging ahead' other than bending over and parting your butt cheeks.
For all the warnings about AI, the truth is that no one knows what kind of world it will produce.
Sure we do. It will be the same world as the internet created, just better.
AI agents could run amok,
unless policed by other AI agents. This is a Red Queen race of diminishing importance. Nobody now greatly freaks out about viruses & 'pop ups' & so forth. It appears that Ransomware payments have fallen below a billion.
forming swarms and turning against their own creators.
Will Cybernet launch all the world's nukes? No. Fuck off.
But a more realistic scenario is that a handful of super-wealthy individuals
none of whom can stand each other
use their control of powerful AI systems to
get assassinated? Basu comes from a country where assassination tempers autocracy. It's all very well saying 'Musk wants to run the world' but Musk is easy to kill.
establish an Orwellian global state
because China has lowered its trousers and is parting its butt cheeks? Fuck off!
—the North Korean model gone global
North Korea doesn't have much in the way of IT or AI
—and end democracy as we know it.
Basu is a Bengali buddhijivi. He wouldn't know democracy if it bit him in the leg.
The sad truth is, democracy is majoritarianism. It means the BJP ruling West Bengal because West Bengal is majority Hindu.
This makes curbing extreme inequality all the more urgent.
China is ruled by a Communist Party. It decided that 'curbing extreme inequality' wasn't urgent at all.
We need a special tax on the wealth of the super-rich, with the proceeds going to those below a certain income.
How else will the US become poor and weak?
As the Roosevelt Institute’s Michael Madowitz recently argued, democracy’s survival depends on our ability to tax extreme wealth.
Democracy survives where no other form of government is possible. Government's have the ability to tax factors in inelastic supply. If 'extreme wealth' is highly mobile or otherwise elastic in supply, it can't be taxed. Basu might profit by learning a little of the subject he teaches.
Oligarchs commanding swarms of AI agents could become more powerful than most governments—a recipe for an authoritarian future.
Oligarchy isn't authoritarian. It tends to be 'Classical Liberal'.
Taxing their fortunes would curb that power
Unless the threat of such taxation causes them to invest in gaining political power. Think of the Koch brothers & the Tea Party & Trump beating Hilary in 2016.
and could help finance less murderous, more useful technologies.
AI isn't murderous. It is useful. No wonder Basu is against it.
No comments:
Post a Comment