In his book Digital Soul: Intelligent Machines and Human Values (isbn: 081334266X), Thomas Georges takes readers on a tour of the ethical, moral, and philosophical problems we may encounter as computer technology advances toward sentience. Georges dives into the field of artificial intelligence, in particular the idea of strong AI, a constructed intelligence capable of self-awareness. But while the tour is informative, many of the questions Georges raises are left unanswered.
We live reasonably comfortably today with expert systems that demonstrate rudimentary weak AI. Amazon.com regulary updates my personalized home page with book suggestions based on what I’ve ordered from them in the past, and what other people with similar tastes have ordered. If you work with databases, this seems far from magical or intelligent. Certainly Amazon has sophisticated algorithms to match your tastes, but in the end these algorithms boil down to relatively straightforward database queries.
It is a big step to go beyond these expert systems to true strong AI, a sentient intelligence that judges things on a combination of emotions and rational calculation. Georges is confident that, gradually, we will make that step, and there are plenty of motivated people working toward that goal, people he introduces in the course of his book.
Confession: As a teenager, I read Asimov, Heinlein, Clarke and Herbert. In the seventies, I watched Lost in Space and obsessively watched Star Trek, both in syndication. The classic space operas of the fifties, like Forbidden Planet, were favorites of mine. And, yes, I’ve attended more than one Star Trek convention. I fell in love with iconic, all-knowing robots and computers of science fiction, and the real world computers I’ve worked with have always seemed, well, a lot dumber.
But all this is true of just about every computer professional of my generation. Unlike the generation that followed us, we loved computers not because we played with them every day, but because we couldn’t. The computers we loved didn’t exist except in Hollywood and in the pages of paperback astro-adventure stories.
The visionary dream shared by the fledgling geeks growing up at the tail end of the Cold War was AI – artifiicial intelligence. Just those two words – artificial intelligence – sounded cool and mysterious. The nerd dream of the seventies was to own a TRS-80; with some clever programming, we could make our own über-brain, capable of passing the Turing test, or at least helping with our math homework or playing a decent game of chess.
And from there it wasn’t to hard to jump ahead in our imaginations to worlds where computers served our every need from cradle to grave. The more ominous horrors of a world run by despotic machine intelligence, or a world where robots own the battlefield, killing without conscience, were not too hard to envision. either. These techno-nightmares have been staples of sci-fi lore for decades.
Are we headed to the more benign world servile machines that truly relieve us from drudgery? Or are we doomed to become servants of mechanized cyber-tyrants? Is strong artificial intelligence ultimately a transformative cultural creation, or a tool to bring about our destruction?
As important as these questions are, Georges does little to answer them. When he does try to explain or answer these questions, he frequently resorts to examples from Star Trekor other televised space sagas. The writers behind these sci-fi dramas undoubtedly strive to address the weighty issues technology may bring to us, but it is hard for me to believe that the deepest thoughts about the nature of intelligence, human or artificial, are found in a weekly drama featuring Commander Data, and a Klingon named Worf. While his enthusiasm for his subject is evident, Georges book suffers from its lack of intellectual rigor. Given the potential downside of strong AI, a deeper examination of the subject than this book offers is needed.
A note to readers: If you have any suggestions for good reading on AI, I’d be interested to hear about it.