I don’t believe in singularity. Computers, no matter how fast they evolve, will not become sentient. Why? Because we do not know what sentient means. I mean we really don’t. We’re an arrogant bunch – we think, therefore we imagine we know everything. Well we don’t. In truth, if we get over ourselves, we’d realise at best we know a tiny, tiny, tiny amount of stuff. We’re like a child in a museum where the rare, fragile exhibits are all out of their cabinets – we stagger about and smash everything up. Our life spans are remarkably short and, worse, our genetic contribution has a half-life just like uranium. In ten generations you will have made a 0.01% genetic contribution. Let’s hope it’s that winning smile not that giant nose, huh?
Now as well as not believing in sentient computers I also think it is utterly foolish to try and predict the future. But here are my AI predictions for 2020:
#1 There will be a demand for Explainable AI
Algorithms are hard math. Chuck in some extra hard math on top and you have AI - something that is so hard to explain and nigh on impossible to visualise. Mmmmm, so why should we worry? Well imagine if you will that you’re smuggling drugs. You see the sniffer dog at border control. Oh no, you’re rumbled. But hold on, you have a great defence – we don’t know how the dog did it, so let’s take the dog apart to find out how?
Trouble is, dogs are racist[1]. No really, it’s true. Dogs don’t start racist, a bit like babies. I mean, they don’t like cats but that’s quite understandable… neither do I. But not because of the colour of their fur, right? The thing is, if the owner – say, the carer/handler/officer – is racist then chances are the dog (who does a fairly decent job of being empathetic to the hand that feeds and pets it) will pick up some bad habits.
AI displays biases. Sometimes it can be racist, sexist, religionist, beardist and all the other ‘-ists’ because data is hardly ever unbiased. I predict this will mean AI will simply have to explain itself, however hard the math is.
#2 We will see an AI pre post-truth checker
Next up? Truth. I love Gartner, not because of all their fancy research (though I do admire that) but because they gave us the phrases ‘trough of disillusionment’, ‘slope of enlightenment’ and the less satisfying ‘plateau of productivity’ (should it not have been the ‘gentle rise of doing stuff’?)[2].
Anyway, AI (and every year, everyone predicts this) will neither cure cancer, stop deforestation, nor end over-fishing. It will probably make racism worse (who let the dogs out?). It will not make us all vegan (though it should) and it will not fix the post-truth world of modern politics. Of course, it won’t fix these things, or will it…?
There is something AI is starting to do really well, and that’s spot fake news. My good friend Professor Chris Reed at the University of Dundee and his team at the Centre for Argument Technology[3] are the world pioneers at this, so my prediction number #2 is that there will be a pre post-truth checker before long that can make the powerful accountable for their woeful misuse of our trust.
#3 AI won’t threaten your job
Here’s a third prediction. I live in a country which, by some economist’s measures, has zero unemployment (oddly the ‘natural’ measure of zero unemployment is actually a touch less than 5%[4], go figure. No, really - you can go figure; have a look at the reference below – it’s a lovely graph and you can play with it). And better than that, it’s true because the US Congressional Budget Office said so (see pre post-truth checker above, and apply when the app is out…).
So with all this nonsense about technology making us all redundant just think this simple thought: Today is the most advanced humanity has ever been. And you’ve got a job, right? We all have (apart from the 5% weirdness statistic above, you’ll be fine soon, you’re just a statistic, so don’t worry). Prediction #3 - Your job is safe, and the robots are not taking you out.
#4 AI will take out lawyers
Especially patent lawyers. This is a personal one, and a prediction I’d actually like to come true. I’ve had a few good ideas in my time (many of us have) and I’m even an ‘inventor’ of a few, but what I want is an AI which will ingest all patents that are extant, let me describe my idea, and tell me if it’s actually an idea, not someone else’s idea, without charging me per bloody minute on the bloody phone.
I’ve known a few of these people over the years and I do so wonder how as a little boy or a little girl sitting in front of their building blocks, their dolls, their paint pots or their crayons then thought ‘You know what I’d like to be when I grow up mummy? A dream destroyer’. Maybe I just answered my own question...
Anyways – Neil Sahota of IBM, the United Nations AI subject matter expert and a Professor at UC Irvine says (indirectly) ’It may even be considered legal malpractice not to use AI one day’[5]. Good on him. Start with the dream destroyers.
#5 AI will save the world
Last one… What do I really, really, really want AI to do?
Well, bring world peace would be nice but unlikely. End famine? Maybe. Make everyone happy? Fat chance. You know what? It’s actually very simple: Break our reliance on carbon.
There is a total of 173,000 terawatts (trillions of watts) of solar energy striking the Earth continuously. That's more than 10,000 times the world's total energy use[6]. AI really can deliver zero carbon. Imagine if we had as much power as we wanted? You could keep warm, grow food using hydroponics at home (plants mostly just need light, heat, CO2 and water – and many of us have a spare loft), use 3D printers to make stuff, and live in luxury unimaginable to our ancestors. AI could allow that. I’m pretty much off grid at home and it didn’t cost nuts money. It’s just the math that is really hard and AI does really, really hard math.
If we had all the energy we wanted, we could replant the forests. And why not fly if the plane burned hydrogen instead of hydrocarbons? You can crack H2 and O2 out of good old H2O with a few of those terawatts, then the vapour trail would be pure water. Instead of dream destroying we’d be cloud making. Then every one of us could have whatever we wanted and stop toddling about smashing up our beautiful museum.
Imagine that?
I have and it’s real.
Be happy, AI will save the world.
______________________________________________________________________
[1] https://www.psychologytoday.com/us/blog/canine-corner/201909/can-dogs-be-racist
[2] https://www.gartner.com/en/research/methodologies/gartner-hype-cycle
[4] https://fred.stlouisfed.org/series/NROU
[5] https://www.forbes.com/sites/cognitiveworld/2019/02/09/will-a-i-put-lawyers-out-of-business/#617bf3fa31f0
[6] https://phys.org/news/2011-10-vast-amounts-solar-energy-earth.html
