Artifical Intelligence: what kind of future will we have?

I’ve been thinking a lot about Artificial Intelligence lately, since it seems this subject keeps coming up on my radar screen. First, early this month, there was the interview with Stephen Hawking in which he warned of the dangers: “AI could end human race” was the headline and Dr. Hawking (a physicist and cosmologist, not a computer scientist) took a very dark view of what AI may do to us in the future.

Next came news that a program had finally passed the annual Turing Test, in which computer programs attempt to fool judges into thinking they are real people. The fact that this year’s winning chatbot – software that imitated the personality of a 13 year old – had finally hoodwinked the judges produced only head shaking from me (I actually said “Can we move on now?”) but the press took the occasion to warn that computers were somehow getting closer to full consciousness. Who knew what dark dangers lay ahead?

But finally I’ve just had the antidote to the doomsdayers, and that is in the form of the excellent new book by Walter Isaacson called “The Innovators”. I could write many blogs about this book – it is an excellent history of computers, software and the internet, and I recommend it highly – but in the closing chapter Isaacson looks ahead, and it is here that he gives a very sane, reasonable and insightful look at Artificial Intelligence.

He mentions that AI has been on the horizon for many years, including an announcement in 1958 that computers that could “walk, talk, see, write, reproduce…” were just 20 years away. It seems that AI is always just 20 years away, including a 2013 story in the NY Times that predicted virtually the same thing.

And I’m sure you remember HAL from the movie 2001: A Space Odyssey. For my generation HAL represented the malevolent super computer that would kill humanity, and it was something we were all watchful for and ready to defend ourselves against.

Yet what super computer did we have, by 2013? It wasn’t HAL. I would point to Watson, the IBM computer that won Jeopardy in 2013 and is an absolute marvel of computing genius and power, both in hardware and software terms. Watson’s performance on Jeopardy, and in other settings, shows that super computers are still under human control, and are really just using brute force computational power to produce results that mimic human behavior. Watson came up with correct answers through the use of fast software that parsed human questions and then examined millions of facts at lightning speed. But it no more understood that it was playing a game than your alarm clock knows that it’s waking you up in the morning when it goes off. Watson imitates consciousness but possesses none.

In his book, Isaacson touches on the realm of human-computer collaboration that I think is at the very forefront of computer work today. I wrote about this in my post "You can be Smarter Than You Think," and Isaacson repeats the story that I also referred to, about the two average chess players that beat grandmasters and supercomputers alike by being the best at collaborating with their tools. It’s a wonderful story, and I think gives us a view of what the future will really be like: not a war between humans and malevolent HALs, but a future rich in progress due to the evolving collaboration between people and computers. A human-machine hybrid, in whatever form that comes about, is sure to be the most brilliant and powerful use of computers that we can hope for.

The most exciting example of this is the work now going on at IBM in which Watson is being teamed up with the world’s most brilliant scientists, not to play Jeopardy or chess, but to search for a cure for cancer. Each side supplies what the other lacks. The scientists bring the creative insights, the leaps of logic, the moral compass and the overriding sense of purpose. Watson has 2 million pages of medical journals, 600,000 pieces of clinical data, 1.5 million patient records, and the power to use it all.

This is the future of Artificial Intelligence that we want, and I’m betting that there’s a good chance that this is the one we’ll have.