Last week, ScienceOmega.com
published an article from Professor Austin Tate, Director of the Artificial Intelligence Applications Institute (AIAI) at the University of Edinburgh. In his piece, Artificial intelligence - from fantasy to reality
, Professor Tate discussed progress made within the field of AI, and some advances that we can expect to see in the future. Although Professor Tate concludes by predicting the advent of autonomous, mechanised colleagues, capable of assisting us throughout our working lives, he begins by commenting on how AI advancements have not occurred at the rapid pace that many had expected.
Essentially, until I am presented with a robot in a sulk, I will continue to refute that intelligent machines exist. Earlier this year, some clever so-and-sos from Drexel University created The Hubos; a mechanised homage to The Beatles. Well I am holding out for The Hubiths, with gladioli included. ‘Waste Electronics are Murder’, ‘Turing Knows I’m Miserable Now’, 'This Charming LAN', etc.
It’s hard to disagree with this statement. If James Cameron had been correct, the remnants of humanity would currently be involved in a battle for survival against an army of soon-to-be Schwarzeneggers. Naturally, I’m glad that Cameron was wrong – I’m in no great hurry to witness the birth of artificial malevolence.
Still, I can’t help but think that the ‘fault’ for this apparent disparity between anticipation and actuality, does not lie with Professor Tate et al. It seems to be a prerequisite of human nature to get ‘carried away’ with an idea. If you don’t believe me, consider Daily Express science reporters, televangelists and English football supporters. I would argue that we are all, to some extent, guilty of this distinctly human tendency. If presented with something new that has been made possible in the here-and-now, it is hard for even the most grounded individual not to wonder about what might become possible in the future. If you’d have shown a plate-carrying robot to me in the mid-1960s, I’d have been hard-pressed not to demand ownership of a bicentennial Robin Williams by the end of the decade.
Some might argue that we could learn a thing or two from our mechanised counterparts. Do androids dream of electric sheep? Well, no – and this makes them good at their jobs. I’m fairly certain that even the smartest of today’s machines is completely focused on the task at hand. I doubt that on May 11th
1997,* Deep Blue became distracted by the lovely beige blazer of Garry Kasparov, but I’m fairly sure that Grandmaster Kasparov probably had a few concerns about having his queen handed to him by a machine. Moreover, Watson didn’t have to expend any computing power on competitiveness, but its Jeopardous opponents certainly did.**
Still, it is not my intention to side with the machines. I am guilty of many things but disloyalty to humanity is not one of them.*** I would argue that capacities such as overstatement, pride, competitiveness and the like, are what separate us from today’s automata. In my – admittedly humble and uninformed – opinion, today’s machines cannot be described as intelligent as they lack the capacity of imagination. Perhaps the smartest examples are able to ask, ‘What if?’, but they are not able to worry
about any of the potential outcomes. They are not capable of hoping
for one thing to happen rather than another.
Essentially, until I am presented with a robot in a sulk, I will continue to refute that intelligent machines exist. Earlier this year, some clever so-and-sos from Drexel University created The Hubos; a mechanised homage to The Beatles. Well I am holding out for The Hubiths, with gladioli included. ‘Waste Electronics are Murder’, ‘Turing Knows I’m Miserable Now’, 'This Charming LAN', etc. Hopes and dreams may well be seen as examples of human frailty, but they also mark us out as intelligent. Until we create machines with the capacity to have artificial ambitions, artificial fears and artificial eccentricities, I don’t think that we will have achieved true artificial intelligence.
I am certainly not saying that it can’t be done. I’m just saying that it hasn’t been done yet. In this week's poll
, we asked whether true AI will be achieved before the end of this century. Ever the optimist, I voted yes. I might not be around to witness it, but I am confident that we’ll have a daydreaming, disobedient machine eventually. My idea of true AI might sound like hell, but really it’s just HAL – frighteningly competent and with its own agenda.
P.S. I want to make it absolutely clear that I have nothing against the fine folks at IBM. They do sterling work and I wish them every success in the future.****
* Not long before the date of Cameron’s predicted Judgement Day
** Tread softly, IBM, because you tread on human dreams
*** I don’t work for IBM
**** I just know which side they’ll choose come Judgement Day