Wednesday, May 22, 2024

OPINION AND ANALYSIS | 27-05-2023 06:47

Things from cyberspace

As well as having nightmares about what would happen to us were a self-programming AI machine to free itself from human control, researchers fear they could be providing evil humans with tools which, like nuclear bombs, would make it easier for them to cause great harm.

Nobody enjoys being treated with contempt by people who are obviously far smarter, so perhaps it is not surprising that some individuals who are accustomed to being regarded as the brightest around have taken to warning us of the dire dangers posed by Artificial Intelligence which, they evidently feel, makes them all look like halfwits. Led by the indefatigable Elon Musk, Apple’s co-founder Steven Wozniak and others who presumably know what they are talking about, they want development of AI, as it is handily known, to be put on hold for at least half a year so they can work out how to stop it from breaking free and going on a rampage. They fear, as the late Stephen Hawking once said, that “the development of full artificial intelligence could spell the end of the human race.” 

There are several theories about how this could happen. Some suspect that, before too long, an AI system could acquire a mind of its own and, after thinking things over, decide it would be in its interest to get rid of creatures who in its view serve no useful purpose. They point out that AI is evolving at breakneck speed, soaking up huge amounts of information and learning how to make use of it, with the result that, by most measures, it is already vastly more “intelligent” than any human being. The idea is that, while it took our species 3.7 billion years to evolve from the “protoplasmal primordial atomic globule” from which, in The Mikado, W.S. Gilbert’s Pooh-Bah traced his ancestry, AI could cover the same ground in a couple of years, leaving us all as far behind as we are from our own microscopically small beginnings. And then? Nobody pretends to know the answer.

Another question bothering people has to do with sentience and consciousness. A Google engineer who told the world that the “large language model” he was tinkering with had become “sentient” got fired for his pains. The consensus is that, unlike earthworms, even the most sophisticated machines cannot really feel, yet alone become aware that they are potentially autonomous creatures and therefore entitled to behave, aggressively or defensively, like all other life forms. Philosophers and scientists are still trying to define consciousness with suitable precision and find out where it comes from, but it would appear that few believe a machine, like a technologically marvellous version of the Golem or the “monster” cobbled together from dead bodies by Victor Frankenstein, could eventually possess it.

As well as having nightmares about what would happen to us were a self-programming AI machine to free itself from human control, researchers fear they could be providing evil humans with tools which, like nuclear bombs, would make it easier for them to cause great harm. For now at any rate, this seems more realistic than fantasies about Terminator-like behemoths running riot in cities or a device that, for what to us would be inscrutable reasons, suddenly pulls the plug on the many services that depend on computers and plunges most of the world into chaos. 

For dictators such as China’s Xi Jinping, AI is a gift from heaven, which is why he is urging a growing army of specialists to ensure that his country gets more of it than any other, including the United States. He and his underlings take it for granted that seizing the commanding heights of the AI business will help China dominate the world by becoming more efficient in just about everything.

It is certainly proving useful when it comes to letting the regime keep a close eye on what is going on in the minds of members of a potentially restive population by checking on what they are interested in and what they say to others. The “social credit system” the Chinese rulers have contrived allows them to keep tabs on everyone and either punish or reward people for what they presumably think. In the United States, many suspect they are also filing away information about North Americans as well as most others throughout the world, which is why they are campaigning vigorously against the tech company Huawei. 

On a more mundane level, AI is already having an impact on many kinds of cultural activities. Not only schoolkids can get easily available “chatbots” to do their homework in seconds, but even doctoral students can save themselves time and effort by asking one to write them a thesis. These applications are also expert at producing plausible fake news items and can churn out convincing videos in which well-known figures appear to be saying or doing extremely unpleasant things. Thanks to AI, would-be tyrants, demagogues and the like will find it much easier to reach their objectives; what would once have been no more than an easily refutable rumour sent out to harm a rival trying to confront them can be made far more persuasive, and can be spread far more widely, that would have been the case until very recently.

If the prophets of AI are right, and there is no reason to think they have got it all wrong, the technological progress which is reshaping most economies is about to go into overdrive. In many places, repetitive tasks, whether manual or white-collar, have already been entrusted to “smart” machines. Others that are rather more demanding will soon be taken over; this is tough news for the middle class, which is getting hollowed out in most developed countries, and confronts democratic politicians with a series of nasty dilemmas. If they try to slow the changes that are taking place they will risk letting their country fall behind others, but speeding them up will be certain to stoke social unrest.

The mere fact that many very influential people think AI could be terrible news for humankind but that, now it has been let out of the box, it cannot be put back, is itself a disturbing phenomenon. What with climate change, the rapid ageing of populations that refuse to reproduce themselves, a death-dealing pandemic which may have originated in a virology lab in China that engaged in what is euphemistically called “gain-of-function research,” the possibility that more nuclear disasters could occur at any moment, intercontinental migration on an unprecedented scale and empire-building by brutal autocrats, people already had more than enough to worry out without needing to be scared by warnings about what thinking machines may have in store for them. 

James Neilson

James Neilson

Former editor of the Buenos Aires Herald (1979-1986).


More in (in spanish)