“Computers are useless; they can only give you answers”. Pablo Picasso
For anyone paying attention – as I know, dear reader, you are – you’ll have no doubt noticed that the debate about artificial intelligence is becoming almost as intense – almost – as the US Presidential campaign.
Every day new articles, new movies, new TV shows, new conferences, new books, appear, warning us that the robots are coming. The latest recent gathering of the global socio-economic elite, in Davos – Woodstock en Piste – majored on the role of AI in the “4th Industrial Revolution” and the impact AI will have on the future of jobs. If there was a Sypder Index Fund tracking AI commentary (traded algorithmically, naturally – pun intended) it would be worth stuffing into your 401k.
There are, of course, many similarities between the AI debate and the political one; both are really, as the WEF recognized, about the future of work; how to get work, how to secure work, what work fundamentally is, in an age where code can do more and more things that humans have traditionally traded for money. And both are examining the nature of artificially; “authenticity” versus “political expediency” is a central tenant of the political race; “wetware” versus “software” is core in the race against the machine.
And just as in the political debate, the loudest, most extreme voices seem, at present, to be capturing most of the oxygen in the room. For Trump and Sanders in politics, read Kurzweil and Musk in AI. For Ray Kurzweil the singularity is near. For Elon Musk, AI represents “our greatest existential threat”.
As per usual your humble correspondent can see both sides of the argument. (In my defense your honor, I present F.Scott Fitzgerald’s famous words, “The test of a first-rate intelligence is the ability to hold two opposing ideas in mind at the same time and still retain the ability to function”). On the one hand it’s not hard to imagine that the offspring of the computer that won Go http://bit.ly/1PTnLAK aren’t going to stop there. On the other hand though, a future where people are oppressed by malign machines seems straight from a back lot in Hollywood or Watford; fun to chew on with some popcorn on a Saturday night but hardly something your angst needs to grind on right now.
If you are increasingly anxious about the rise of machine intelligence though I’d strongly recommend the recent book What to Think About Machines That Think http://amzn.to/1QEAPg6 edited by John Brockman, Editor and Publisher of the very cool http://edge.org/ Brockman has rounded up essays by a lot of top wetware including Paul Saffo, Tim O’Reilly, Kevin Kelly, Nick Bostrom, and Esther Dyson, amongst many others, who all take different cracks at figuring out quite how worried/excited/blasé we should be. If there’s a consensus amongst the thinkers about what to think about machines that think it’s pretty hard to discern. Better software than what I have on board would be needed to figure that out. There are some, like Bruce Schneier from Harvard Law School, who wonder what happens when a omputer (not its operator) breaks the law. Others, like Josh Bongard from the University of Vermont, suggest that when “machines are commanded to “survive, reproduce, and improve in the best way possible” they will probably give us humans a very short window to relish that insight”. But then again, others feel aligned with Steven Pinker from the Department of Psychology at Harvard, who writes “My own view is that the current fears of computers running amok are a waste of emotional energy – that the scenario (i.e. the rise of AI) is closer to the Y2K bug than the Manhattan Project”.
Which brings me full circle back to that primo piece of wetware (not featured in Brockman’s book) Pablo Picasso. In 1968, when he made the remark above, computing was – as seen from 2016 – in its infancy. But it was powerful enough to fire the imagination of many who wondered where software would take us and what it would be like when we got there. Arthur C. Clarke was worrying about HAL. Philip K. Dick was dreaming of electric sheep. Michael Moorcock was already leaping ahead to speculate about the final program (US spelling!). It wasn’t hard, even then, to extrapolate that software would get cleverer and cleverer and one day get cleverer than us. Of course the Turing Test was already almost 20 years old by then, long enough for the kings of the swingers to worry how long they’d be the jungle VIPs (Disney’s The Jungle Book came out in 1967 ...)
Picasso though was less impressed and less concerned. His, at first gnomic, but on reflection, devastatingly profound statement, was spot on then, but even more spot on – even with the technological advances of the last 50 years – today. Computers have gotten great at giving us answers; ask Siri “what does the fox say” and she’ll (!) correctly answer “fraka-kaka-kaka-kaka-kow”. Waze will tell us to avoid Route 17 because there’s just been a fender bender. Zapier will automatically create a Google Calendar event from an Evernote reminder. IFTTT will tell your family when you’re on the way home from work. But these are still all answers. We – the wetware – are still thinking of the questions. Even on the very far edge of the new frontier – the aforementioned game of GO – the real VIPs are humble and understated in their claims for where things stand; “It’s (sci-fi AI) very, very far in the future from the kinds of things we’re currently dealing with, which is playing Pong on Atari”, Demis Hassabis told the FT last year http://on.ft.com/1P3i59q. This nice compilation of quotes http://read.bi/1GTtZge, rounded up by Business Insider, also injects a dose of healthy skepticism from those really in the know.
Now I wouldn’t go quite as far as Old Pablo in saying computers are useless; I quite like my iMac, iPad, PC, Apple Watch, E20, TV, car, etc etc (computers one and all). But I do think that Pablo Diego José Francisco de Paula Juan Nepomuceno María de los Remedios Cipriano de la Santísima Trinidad Martyr Patricio Clito Ruíz y Picasso (would a computer think of that?) was onto something then and is still onto something now. We – not software – are the future of questions and questions remain more important than answers. Questions come first. Every answer begets a question.
Questions – curiosity –are/is (I would argue) the central defining characteristic of intelligence (as it is manifested in our human form). From our first words http://bit.ly/20ErHNk to our first steps to our first journey it is intrinsic to our very being to want to know who what why or where. Nobody tells us to ask questions. No parent, or teacher, or TV show, or social media feed tells us – programs us – to want to know “zup?” We just do. I wonder why? C’est une bonne question...
When computers start asking questions – “just what do you think you’re doing Dave?” – then I guess we can start worrying. But that – it seems to me – is as far as away in the future as it was 50 years ago.
Please note that this is article was written by me, Ben Pring, not by Automated Insights or Narrative Science.