The media’s been captivated this week by the story of Geoffrey Hinton, the supposed “godfather of AI” who left Google to speak more freely about the supposed threat he sees artificial intelligence posing to humanity — and which he ultimately helped to advance. You can see why it’s such an attractive story, but the focus on Hinton and his framing of the problems with AI continues to distract us from the real conversations we should be having that aren’t focused on some imagined future, but on the harms these technologies are causing in the present.
For Hinton, the threat of AI is a future problem. It’s not about how AI will increase the power of employers over employees, how it will be wielded against marginalized communities, its potential environmental impacts, and other serious concerns that will affect a lot of people outside the circles of wealthy technologists and executives. Instead, Hinton’s focus is on the fantasy that AI is on the cusp of becoming more intelligent than humans, and will then have the capacity to trick and manipulate us into doing its bidding.
In an interview with The Guardian, Hinton explained
The reason I’m not that optimistic is that I don’t know any examples of more intelligent things being controlled by less intelligent things. You need to imagine something that is more intelligent than us by the same degree that we are more intelligent than a frog. It’s all very well to say: ‘Well, don’t connect them to the internet,’ but as long as they’re talking to us, they can make us do things.
Let’s be very clear here: this is not the reality of this technology. Just because ChatGPT can churn out some convincing text does not mean it’s anywhere near having the ability to think, let alone reach the level of human intelligence.
This is the dream of many of the tech folks behind this technology. Even though they claim they don’t want it and that it poses a threat to humanity, they’re still actively working toward and dreaming of a day where artificial general intelligence (AGI) becomes a reality — where the computers can think for themselves. But just because they can dream it, doesn’t mean it’s anywhere near being achieved.