Apologies to Drake for inelegantly stealing his line but I’m in Canada this week, speaking at the iX Symposium at the Société des Arts Technologiques and I’ve become very interested in the concepts of artificial intelligence and how mainstream the expression has become. The world of the internet of things in which I mostly operate has been shoved aside by pundits and the press in favour of the ‘flavour of the year’AR/VR/AI/cryptocurrency. In this new wave of techno babble, some trends are clear:
- Wilful ignorance of the experiential and hardware limitations. How many headsets can you ship, how much do they cost, when are you using them and for who? seem to be questions no one seems to be interested in.
- Misunderstanding of simple computing principles. Most people use ‘AI’ when they just mean ‘computers’ or ‘maths’. More on that below.
- Misunderstanding of the hardware realities of computing principles. No, no and for the last time, no, you can’t put (X) on the blockchain, especially if there’s a hardware component to (X) which implies a supply chain, which implies people. Just forget it. You can’t track a fruit from birth and you don’t want to track child exploitation in the fashion world. So there.
I’m rereading The Golden Notebook and was struck by a line early on:
‘What’s wrong with living emotionally from hand-to-mouth in a world that’s changing at fast as it is?’.
Doris Lessing wrote this in 1962 but it could have been written in 2018. In our recurring feeling of being ‘in a frenzy’ all the time, we’ve, in fact, made little progress in utilising new technologies for socially useful purposes.
I posit that Artificial Intelligence could easily be described as:
- Computing power doing stuff faster than I can with a piece of paper. (Maths)
- Computing power utilised to make me spend more money than I have. (See all travel ticket, hotel, or insurance purchases ever.)
- Computing power utilised to confuse me into spending more money than I’d like. (See Amazon who isn’t actually the cheapest place for anything.)
- Computing power utilised to reduce the amount of time spent doing something boring, making me spend more time in meetings that don’t need to happen. See all office jobs ever.
- Computing power that creates a new category of boring jobs in the moderation, data tagging, data verification sectors. (See police officers that can’t get a computer to recognise the difference between pornography or dunes or failed use of IBM Watson in cancer identification or all the people Facebook has to hire in its security group).
- Computing power that moves our admin burden from paper to online but doesn’t remove it. (Read about UK nurses who can’t do their jobs because of the admin load)
- Computing power that barely understands us but pretends to. (See Alexa that orders stuff on its own or starts randomly recording.)
Great work everyone. If computing power isn’t there to help us become better societies, then why exactly are we using it? Where are we going? What are we not designing instead? What are we avoiding because it’s supposedly ‘too complicated’. A lack of ambition shouldn’t be confused with a lack of technical capabilities. But if we’re not ambitious about what we want from our computers, we have to ask ourselves who we’re protecting by that cowardice. The rich? The powerful? The establishment?
These are some of the topics I hope we’ll talk about this week, because I really had had enough of us talking about AI without pointing out what exactly it is, and crucially, what it isn’t.