EAGLE-EYED readers will recognise this as our second special issue of the year on artificial intelligence. It is unusual for us to return to a topic so soon. But AI is moving so fast and attracting so much comment, from the insightful to the hysterical, that it is important to offer a clear-eyed look at the risks and opportunities it brings.
The first thing to say is that there is more to AI than the new generation of chatbots that have sparked the recent hubbub (see “How does ChatGPT work and do AI-powered chatbots “think” like us?”). Different forms of the technology have already solved some serious scientific problems. Researchers have applied AI to everything from protein folding, which is revolutionising biology and drug development, to the quest for commercially viable nuclear fusion (see “The biggest scientific challenges that AI is already helping to crack”).
Advertisement
That isn’t to say generative AI, which creates text and images in response to human prompts, isn’t useful. Forecasts suggest a general productivity boost from these systems could raise annual global GDP by 7 per cent (see “What generative AI really means for the economy, jobs and education”). If distributed equally, and not funnelled exclusively into the hands of shareholders, that could lift millions out of poverty and improve quality of life. We might all put generative AI to good use to save time on day-to-day tasks we might otherwise find tricky or tedious (see “How to use AI to make your life simpler, cheaper and more productive”).
But there are risks, and here we must think clearly with regard to what to worry about. Rather than focusing on vague warnings of existential threats from a hypothetical super-intelligent AI, or even a conscious one (see “Can AI ever become conscious and how would we know if that happens?”), we should pay more attention to extant problems.
One immediate risk in danger of being overlooked is AI’s ability to amplify the spread of false information (see “Forget human extinction – these are the real risks posed by AI today”). Distortion or fabrication of facts by AI could be accidental, due to sloppy safeguards, or it could come in the form of deliberate propaganda. In any case, with elections around the corner in both the UK and the US, surely the most pressing concern is AI’s capacity to supercharge the distortion of democratic processes. The worry is that there seems to be precious little in place to combat the threat.
