In the 1990s the internet broke out of the Ivory Tower and started heading for the High Street and our living rooms. There was a mega buzz about the technology’s progressive, transformative powers. Silicon Valley and its acolytes started to promote the idea of “innovation” as a fundamental right.
Eventually it dawned on people Silicon Valley’s notion of “innovation” was just a less pointed, less worrying, way of saying, “experiment”. We were the guinea pigs and the aim was to see if we could be socially engineered into milch cows.
Through slick marketing and effective lobbying, politicians in the liberal democracies were persuaded not to “step in too soon”, not to “interfere” with the experiment. They were anaesthetised by words like “human rights”, “free speech” and“democracy”. However, Mahatma Ghandi and John Stuart Mill have now unequivocally left the building. Around the edges there persists a doughty band of activists, mainly lawyers, academics and techies who cling to their original hopes for the technology but the key word there is “edges”.
I don’t doubt the likes of Mark Zuckerberg and Tim Cook would be delighted if they thought they were helping advance any number of indisputably worthy objectives but on a list of one to a hundred where do you think such ideas would stand as discrete corporate objectives? Any claim guys like these might once have had to be leaders or prophets of a new and better age now seem ludicrous. We obviously inhaled too deeply or were just straight-up conned.
True enough the new businesses the internet spawned would not have succeeded if they had not come up with products they could convince us we needed. We would only buy and continue to use them if we could see an advantage or benefit from so doing. The internet and associated digital technologies have also engendered all manner of improvements in public administration and been a boost to economic growth, but look at the avoidable price too many people have had to pay, including too many children.
Tech could have acted voluntarily to anticipate and prevent bad things happening in the first place or, once discovered, acted swiftly to put them right. They didn’t. Now legislatures everywhere are trying to retrofit guardrails.
Is AI heading the same way?
The current discussions on AI echo many aspects of the early debates on the internet. Once more politicians are being urged not to “step in too soon”. See also the recurring rows reported in the media about battles between idealists in the AI space and the guys putting up the money. If recent history is any guide…… see above.
Some say the genie is already out of the bottle. It’s already too late to regulate. That’s a counsel of despair. Cui bono?
What is true is, unlike the early days of the internet, when western-style democracies had a head start because, inter alia, that was where the tech developed and first took off, today tech knowledge and expertise is much more evenly distributed around the world which means it is in the hands of some rum characters who are every bit as smart as “us”, even if they pursue very different agendas. They can see the potential for AI to help maintain their bloody grip on power.
This does not mean the liberal democracies should abandon the attempt to develop a robust regulatory regime which will help ensure, as best we can, that we do not fall foul of some of the obvious worries being expressed about AI, of which misuse for political ends is but one.
The UK, the EU and maybe some others could be persuaded to make a joint effort to set standards and work out defences against malign uses of AI. It would be nice to think the USA could be part of it but writing today who can feel confident about anything tech-related coming out of there?
Will, could, AI develop to a point where it sees humans as a threat to “the mission” then proceed to end the reign of homo sapiens? Or will it free us all from daily strife, enabling us to stroll through sylvan glades composing poetry and playing pan pipes?
The answer to those questions should be
“Don’t be ridiculous. Both ideas are absurd”
But the very fact eggheads with considerable knowledge of these matters think at least the first question might not be ridiculous, should give us pause for thought. Thus, precisely because there is room for doubt and the consequences of getting it wrong are so apocalyptic, our politicians need to roll their sleeves up and get stuck in.
We certainly cannot rely on the self-restraint of smart techies looking to displace Jeff Bezos and Elon Musk as the world’s richest pop stars nor can we put our faith in anonymous nut jobs working for Kim Jong Un.
People sneer at Governments, denounce their oft observed incompetence, but if the only alternative they can come up with is trusting people who stand to become super rich by directing AI one way rather than the other then they have really not come up with an answer at all.
AI’s capacity to facilitate the creation of pseudo csam, the way Large Language Models have introduced algorithmic bias through the indiscriminate scraping of sources, the way published authors and other creatives are being ripped off, has given us the merest hint of possible downsides to AI, and don’t get me started on AI’s thirst for energy to power the necessary computing processes. But then there’s the advances in medicine achieved by, for example, DeepMind’s AlphaFold and Benevolent AI’s work on new drugs and treatments.
AI is not a single thing. My hunch is individual applications which incorporate different forms of AI will produce huge benefits in several areas but I cannot ignore those better informed voices who point to the risks. Governments have to engage, in the public interest. Nobody else can.
AI to help children
A few weeks ago I went to Sofia to speak at a conference organized by Terre des hommes. It looked at how AI might be used in wholly beneficial ways to help children. I will do a short report on it in my next blog. It will appear very soon.
Desiderata 