AI and the ‘every’ problem

A riled Chris Yapp gazes at the hype surrounding AI and suggests that the headline writer’s vision of 2050 may be completely and utterly wrong.  

If there is one thing that constantly amuses me - or riles me - in equal measure it is the sloppy use of: ‘every...’

When I see a headline saying: ‘The series everyone is talking about’, I think: ‘I’m not, I’ve never heard of it and I pity the deluded author.’

There are many lists of ‘100 books everyone should read’ or the ‘50 films everyone should see’. The idea that someone thinks that I am a nobody because I don’t buy into their cultural norms is snobbery at its worst. I amuse myself at the size of the overlap (personally around 20%).

As with the ‘all swans are white’ problem, it only takes one black swan to falsify the claim.

Progress in AI and machine learning is certainly impressive and I am optimistic that, in the long run, the benefits can, and will, accrue to us mere mortals. However, I do wish people would stop making claims that: ‘By 2050 AI will outperform people in every domain’

What exactly does that claim assert? The problem with sloppy claims like this is that it raises fears about dystopias that get in the way of rational debate. Techno perfect futures contrast with the real-world experience of all of us. Of course, the internet and WWW have brought huge benefits to society and the economy, but the concerns over fake news, cybersecurity and online bullying are real.

Step into the time machine

The Year is 2050. Every teenage boy’s first date will go perfectly. On his iPhone40 there will be an app powered by big data and algorithms so powerful he will know ‘what women want’. Or maybe we won’t need to date, as AI will be able to pick your partner for you better than you can. Imagine the 16-year-old who is told that they are LGBT when they haven’t yet realised that. No-one will ever have an affair or divorce because the AI matching will be vastly better than any human decision making.

I hope that you accept that this is absurd. If you don’t I suggest that you get out more.

So, what exactly would be a claim in this area that might be a more realistic proposition?

Take the claims around super intelligent general AIs. By 2050, an AI will have greater capacity than all humans put together.

So, let’s take this 10bn human equivalent AGI. How long would it take to unify general relativity and quantum mechanics to create the Grand Unified Theory? It’s been a goal of some of the finest human minds for over 100 years. I’ve listened to talks on this that suggest that it’ll probably be done on a wet Thursday. I am of course assuming humans haven’t accomplished it fully by 2050, by the way.

I’m happy that the laws of robotics, the ethical stance is being debated seriously. The debate on the nature of society and economy harnessing these developments we wish to build is maturing quickly and well. The world will be very different.

I suspect that Red Dwarf and The Hitchhikers Guide to the Galaxy will be closer to the truth than some of the Industry 4.0 tech determinism that I’ve read.

There was a interesting programme on Radio 4 about the making of ‘Ex Machina’, a film that has both utopian and dystopian elements in an interesting balance. I enjoyed the film more after listening to the debates about the background to the film than I think I would have without listening to some of the dilemmas they faced in telling the story.

As I have argued before on this blog, technologies are defined as much by their limits as they are by their functionality. We should be celebrating the progress that is being made. Can we just please drop the ‘every’ hype?

About the author
Chris is a technology and policy futurologist. Chris has been in the IT industry since 1980. His roles have spanned Honeywell, ICL, HP, Microsoft and Capgemini. He is a Fellow of the BCS and a Fellow of the RSA.

See all posts by Chris Yapp
December 2017

Search this blog