The problem isn't with evil developers, or evil companies, the problem is that there is no way, even for the tech companies that create it, to control AI. Many of the tech bros developing it are telling you exactly that -- loudly. It's probably a good idea to listen.
It's like social media -- the algorithms were developed to help bring you content you like. What's not to like about that? The algorithms weren't evil in intent (in the vast majority of cases), but those well-meaning algorithms have been heavily leveraged to create huge problems (political influence, kids committing suicide, massive divisions, false information believed by large segments of the population, etc). Social media companies could improve it a bit -- but the reality is that anybody who knows what they are talking about will tell you that they can't really control the algorithms -- and those algorithms are heavily manipulated, constantly, on a daily basis, for all kinds of purposes, many of them harmful or outright evil.
The potential issues from AI are easily 100 times worse than social media. Even among well-meaning developers, and those who use AI for admirable purposes, there is huge risk of unintended consequences --but certainly 1000 times more risk among those who want to use it for harm (e.g. massive financial hacking moving at light speed). If a country chooses to use it for military decisions, for example, what could go wrong? (spoiler alert: a lot).
If you use AI at all, you'll see that it is very frequently dead wrong -- but it always sounds authoritatively irreproachable in its argument. When you interact with it to tell it where it's wrong, it will say something like "that's a remarkably astute observation" and then correct that error, and you can navigate it to some brilliant results by error-correcting it -- but that's only if you know enough to recognize when it's wrong.
There has never been technology this great -- or this dangerous -- or anywhere close. And no other technology where the very people developing it, even in the early stages of developing it where they are actively competing to develop the best platform, are shouting that it is a massive risk. Everyone should listen to them and do whatever they can to help ensure actions are taken to control it.
www.bbc.co.uk
