Do the benefits of AI outweigh the risks?

I was kindly invited to a Raspberry Pi Foundation offsite to debate ‘Do the benefits of AI outweigh the risks?’ Here’s the short statement I shared:

Sometimes the role of an ethicist is to ask distinguishing questions. One such question is ‘for whom?’. Do the benefits of AI outweigh the risks? Well, the benefits for whom? The risks to whom? These two questions will probably have very different answers, right? The benefits and burdens of AI, as with almost every other innovation, won’t fall equally.

Technologies are always imprinted with values. This goes for even the crudest objects, things we don’t even think of as technologies. Think of razor wire. Razor wire is a shockingly opinionated object: it argues that someone’s right to private property is so important that we should injure anyone who violates that right.

AI people love talking about the value alignment problem: the idea that if we create a superintelligence we’d better make sure it holds the same things dear that we do, otherwise it might destroy them. But what happens before that? What values can we see imprinted within the AI systems we’re building today? When I look at modern AI, I see plausibility trumping truth. I see speed galloping ahead of safety. I see disruption hailed as inevitable, as destiny. Now, these may not be intentional design decisions but nevertheless, they have real-world impact. And the choice not to engage with the values and ethics of our technology is itself an ethical choice: an affirmation of the status quo, a vote to stay on our current heading.

AI could well be the largest force multiplier we’ve ever made. But we already feel society’s invisible, systemic forces acutely. Some people are elevated and empowered by these forces. Some are crushed. If we keep fostering the same values in technology that we do today, then I think these injustices will only increase. People who lack power today will end up further robbed of their autonomy and dignity. Entire creative classes may also find themselves dragged down by the technological undercurrents. It’s not hard to imagine a world in which the tech giants collect handsome royalties for their AI’s creations, while painters and novelists have to collect the recycling.

But it doesn’t have to be this way. Technology doesn’t hold the reins. We do. If we can subvert the default values of today’s tech sector and instead build AIs that prioritise compassion, justice, respect then yes, I think the benefits of AI will far outweigh the risks. How we achieve this within the confines of growth and profit is perhaps another question.

Image by Alan Warburton / © BBC / Better Images of AI / Plant / CC-BY 4.0