鈥淎rtificial intelligence is already here,鈥� says , political scientist of the Oxford Martin School at the University of Oxford. 鈥淭he real question is: where are we on the spectrum? How capable are the autonomous brains we鈥檝e created?鈥�
I鈥檓 sitting on a hay bale in a large white tent waiting to find out. Kutarna is debating 鈥渢he road to artificial superintelligence鈥� with AI consultant , Matthew Taylor of the RSA and John Lloyd, producer of TV show QI and former housemate of author Douglas Adams. And they couldn鈥檛 have picked a nicer setting 鈥� a cool tent on a hot summer鈥檚 day, in a field in Oxfordshire, during the .
Before we can figure out how intelligent computers are, or can be, we need to decide what we mean by intelligence. When it comes to artificial intelligence, we tend to keep raising the bar, says Kurtana. We might have thought that it takes intelligence to win a game of chess but the world鈥檚 best chess player was beaten by a computer in 1997. Now, anyone can download a cellphone game that can beat them at chess.
Advertisement
Game on
We鈥檝e also seen AI beat humans in games like Jeopardy! and, more recently, at the ancient and even more complex Chinese game of Go.
But that doesn鈥檛 mean that computers have intelligence. 鈥淒on鈥檛 believe the news,鈥� says Smith. 鈥淭here have been no recent breakthroughs in artificial intelligence.鈥� While computing speeds have accelerated and now exceed those of our own minds, intelligence is more than mere computing, he says.
To fully understand intelligence, we need to understand how the brain works, and, ultimately, human consciousness. And we are a long way away from being able to do that, says Lloyd. As he advised rapt listeners, 鈥淒on鈥檛 buy a book called How the Brain Works, because we just don鈥檛 know. We don鈥檛 know the slightest thing about consciousness, says Lloyd.
Smith isn鈥檛 even sure consciousness exists. 鈥淚t may be a combination of memory delay and imagination,鈥� he says. 鈥淚 think consciousness is an illusion 鈥� when we stopped believing in a spirit, consciousness became a substitute.鈥�
Intelligent machines?
And even if we could create an intelligent machine, would we make good use of it? Our brains are capable of incredible things, but most of us use them in mundane ways to live boring lives, says Lloyd. 鈥淭he more sophisticated the device, the more banal the information shared over it,鈥� says Lloyd.
鈥淎 phone now has more computing power than a human mind, yet we use it to share selfies of our bottoms with each other.鈥�
More worrying is the idea that should we develop AIs, we will expect them to solve all of our problems, like a god. 鈥淚 think we live in a world where technology has become a religion,鈥� says Taylor. 鈥淲e鈥檝e now gone back to the Medievalist notion that the world is whatever God thinks it should be, but God has become technology.鈥�
Part of this is down to the politics behind the technology, says Taylor. 鈥淪ilicon Valley has a view of problems called solutionism, and it is full of the idea that all the problems in the world can be solved with technology,鈥� he says.
Kutarna agrees. 鈥淲e expect technology to keep making things faster and smaller, until eventually there are no more problems and we live in a techno-utopia,鈥� he says.
Tech to serve us
鈥淚t鈥檚 like saying that technology is all-seeing and all-loving and, hey presto, in 30 years鈥� time the technology will be in charge,鈥� says Taylor. 鈥淚t鈥檚 bollocks 鈥� the role of technology is to serve us, not the other way around.鈥�
At any rate, we won鈥檛 be able to give computers human intelligence unless we learn how to instil artificial empathy.
Empathy and emotion is particularly vital to the way our brains work, reminds Taylor. We start forming emotional memories long before we form memories for events, for instance, and these memories shape our behaviour and the way we interact with people for the rest of our lives. No computer can interact with a person in a way that feels natural to most of us.
Perhaps that鈥檚 what makes us special, says Taylor. 鈥淲e need to fundamentally connect to our humanism, and fall back in love with ourselves, he says. 鈥淲e need to get back to appreciating what an amazing species we are.鈥�
One of the soundest reminders of human brilliance is comedy, says Lloyd, which artificial intelligence has never been able to achieve. 鈥淓very year, Richard Wiseman makes a list of jokes written by a computer,鈥� he says. 鈥淎nd by god they鈥檙e shit.鈥�
