AI is not dangerous yet, but it is becoming more uncomfortable

AI
Image credit: Valery Brozhinsky / Shutterstock.com

We have reached the point where we groan at press releases that claim that [insert name of company or product] has AI ‘built in’. We also mentally discard them as probably not serious, certainly not dangerous.

Except AI is on the verge of becoming if not dangerous in itself, extremely powerful in certain uncomfortable areas.

Just last week we reported that a group of researchers had boycotted South Korean university KAIST over its plans to partner to create killer bots, which seemed to herald the first steps towards the weaponization of AI. They have now reversed that decision because KAIST agreed not to develop killer robots after all.

If you have a company and you are into AI, now is the time to be raising money. SenseTime, just such a company, recently raised $600 million, doubling its value. SenseTime provides a key component for the ‘world’s biggest surveillance system’ – China. It provides the facial recognition software that will help China build its controversial (at least to the likes of us) social credit system. The Government plans to amass so much data on its citizens that it will be the one that decides where their children go to school, who they should be friends with and so on.

AI may not be at the ‘robots taking over earth’ stage but as the fast look-up version becomes ever more sophisticated, it brings more power to the powerful.

While that is surely not a good thing, it is also inevitable.

It is also a trend not helped by the growing idea that Silicon Valley exists in its very own bubble, which is removed from the real world by a good parsec or two. As John Tanner wrote a while ago:

A big part of the problem […] is that Silicon Valley is a bubble unto itself that has become so isolated from the way the real world works – and what potential customers might want or find useful – that they’re not only unable to relate to it, they also have no idea why everyone’s so upset […] Why wouldn’t people think replacing their jobs with AI is totally cool? Why would anyone think a virtual assistant is like wiretapping your house? It’s Amazon and Google doing it, not the government, so it’s not like 1984, right?

The thing is that is how Silicon Valley does think. If it is cool technology, let’s roll it out.

It may be that Facebook has burst that bubble, and it may be that some form of regulation is about to descend on internet companies. It may just be that if that happens, then the economics that makes the internet work (and which have undoubtedly been thoroughly abused) will need to be rethought.

Whatever the outcome of those debates, one thing is clear – AI will underpin surveillance, security and strike capabilities within a few short years.

And that means that every press release we get which contains AI now makes for slightly uncomfortable reading.

1 Comment

  1. I think if governments bribe citizens with a Universal Basic Income (UBI), watch how easy it will be for the average person to just take their “Soma-UBI” and raise a quite impenetrable wall between the “fear of AI” and their minds. (Paraphrased with Huxley’s vision)

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.