artificial intelligence | The Casual Blog

April 16, 2011

Artificial intelligence, The Most Human Human, and a walk at Crabtree Creek

I’m still feeling odd and shaken by Watson’s victory over the human champions of Jeopardy. It is truly awe-inspiring that our greatest software engineers have created a program that behaves in some ways like human intelligence, but, at least as far as raw knowledge and research is concerned, much better. What’s unsettling to me is not just the economic implications of this new generation of artificial intelligence, but also the moral/ ethical ones. The new AI is getting very good at the sort of intelligence that we’ve always considered the crowning and distinctive feature of the human race. It’s now clear that our destiny is not to be the most intelligent beings in the universe. So then, what is it? What do we do?

Like Ken Jennings, the former Jeopardy champ who acknowledged defeat with becoming humor and grace, I also welcome our new computer overlords. They already are making daily life better in some ways. I recently had an encounter by telephone with a computer dealing with a travel reservation problem that performed substantially more efficiently than some humans. Later, when I found myself in a phone conversation with a human on another routine matter (activating a new credit card), as I tried to understand the person’s accent and waited for a sales pitch to conclude, I thought affectionately and longingly of my dear computer. Our computers are getting to be good clerks, and I expect they will soon be good scientists, doctors, and lawyers. The trend is clear.

So, what’s left to aspire to? I’ve been reading The Most Human Human: What Talking to Computers Teaches Us About What It Means to Be Alive, by Brian Christian (in the Kindle edition). Christian treats the implications of artificial intelligence for humanity in a lively way. He takes off from his participation in an annual competition for the Loebner Prize, which involves the Turing test. Human judges converse (via networked computers) with both humans and AI programs, and the annual prize goes to the program that does the best job of convincing the human judges that it is human. There’s also a prize for the humans that talk with the judges, and Christian competed for this: the most human human.

The success of some programs points up how much of human discourse is routine and predictable. The weaknesses of the program show that there’s still some human behavior that is creative and (so far) unpredictable. Christian uses the Loebner Prize as a jumping off point for an entertaining, though jumpy and digressive, introduction to AI and its philosophical implications.

As Christian notes, humans are distinguishable from programs with respect to the physical world. We have bodies that are, in their non-rational way, intelligent. Our cells are connected with each other, and our individual bodies are connected to other humans, other species, and the earth, the air, and the Sun. We depend on all these connections. As obvious as this sounds, we still as a race we have trouble keeping in mind our connection to physical reality.

This may be part of the explanation for the right-wing attack on the environment. In the NY Times today, there’s a front page story by with the headline Push in States to Deregulate the Environment. As the story notes, Republicans in North Carolina are proposing enormous cuts to the budget of the NC Department of Environment and Natural Resources. I’d been monitoring the NC story, but learned that the same thing is happening in other states. At the same time, the Republicans in Congress are looking to cut the EPA and gear back environmental regulation.

What is the matter with these people! In the midst of ongoing extinction of entire species and global warming that threatens entire human populations, there should be no debate about the need for new and more effective conservation of natural resources. I have assumed that the opponents of science and environmental regulation are either unbelievably ignorant or unbelievably greedy and cynical, concerned only with immediate short-term gain and without concern for future generations or the earth itself. But to give them a slight benefit of a doubt, perhaps the problem is that they’ve really lost touch with their bodies and the earth.

Anyhow, in addition to enjoying new AI tools, I’ve been trying to make a point to spend a little more time outside. This morning I went to the swamp area of Crabtree Creek off Raleigh Boulevard. It was overcast and windy, but there were hundreds of birds singing and flying. I walked along the boardwalk at the side of the swamp. There were six great egrets, several great blue herons, and numerous swifts and swallows. I saw a black-and-white warbler, a phoebe, a yellow-rumped warbler, cedar waxwings, and heard, along with the common residents, a parula, and a hooded warbler, as well as a handful of songs I couldn’t identify. I love the spring migration season. It’s good to just clear out the mind and just look and listen.

February 19, 2011

Watson, human games, and the twilight of the gods

Sally and I flew out to Telluride, CO yesterday for a late winter ski adventure. On the flight from Raleigh were our good friend Charles and Chuck, and we looked forward to meeting up with Gabe and Jocelyn. The flights took off on time and progressed in an orderly way. I made some progress getting through back issues of The New Yorker, Scientific American and Golf Digest, listened to Mozart and Debussy. And as often happens when I travel at 35,000 feet, I found myself in a contemplative mood. As Garrison Keillor says of his private eye character: one man’s still trying to find the answer to life’s eternal questions.

What is the meaning of play? When humans have taken care of the essentials — food, clothing, shelter, sex — it is a large part of what they do. I suspect the same is true of all animals, based on the birds, squirrels, fish, cats, dogs, and other creatures I’ve observed. They all love to play. Children love to play. Put a random group of four-year olds together and a game will almost always develop.

The games people play vary widely according to their age, traditions, fitness, intelligence, financial resources, and moxy. Some like skiing, some prefer bowing. Some go for chess, and others like checkers. The arts are unquestionably a form of play; we even refer to musical activity as playing music. A lot of our verbal activity has little to do with survival and qualifies as mostly play.

Smarter-than-normal people tend to like games requiring a good memory and a quick tongue, and to view success in those games as a badge of honor. Before this week, we mostly felt confident that, whatever our weaknesses and failings, we were superior to all other known beings at such activities. After Watson’s triumphant performance at Jeopardy this week, that’s over.

I didn’t see the entire three Jeopardy sessions, but I saw enough to get the idea. The gifted engineers at IBM have taken artificial intelligence to a whole new level. (By the way, congratulations, guys.) Watson has incredible facility with language and memory. The humans never had a chance. I was reminded of the song about John Henry, the great swinger of the hammer, who drove himself to death but couldn’t beat the machine. (Bruce Springstein does a great high-energy version of the song.). Admittedly, Watson’s abilities don’t extend to the entire range of human intelligence. For example, it isn’t good at creative reasoning — yet. But the day when it will be considered hopelessly romantic to think that humans could be more intelligent than machines is well within view.

So where does that leave us as a species? Consciously or subconsciously, we justify a lot of atrocities on the theory that we’re superior as a species to all others, Could Watson make us just a bit more humble? Could it inspire a bit of self-examination? If intelligence isn’t our greatest achievement, if compared to our computers we’re not really very bright, perhaps we’ll come to view our most important defining characteristics as other human qualities, like love and kindness. What if we consciously cultivated those qualities?

August 1, 2009

Do humans really control computers, or vice versa?

Computers are the smartest things in the world, and they are throughly embedded in our lives. The good news is they do amazing things. The bad, or at least humbling, news is we will never again be the most powerful intellects on the planet. For better or worse, computer intelligence is changing what it means to be human.

I was surprised that the NY Times published John Markoff’s piece last week on artificial intelligence under the headline, “Scientists Worry that Computers May Outsmart Man.” http://tiny.cc/sDCN8 In 1997, IBM’s Deep Blue chess computer beat then-world champion Gary Kasparov. This was, for me, that was a watershed — strong proof that the era of human intelligence as the dominating force on Earth was drawing to a close.

Today, it’s obvious that computers not only can “outsmart man,” but even in a below average laptop computer is much, much stronger at certain types of information processing than any living human. We can’t even come close to competing with them, any more than we can fly like falcons, run like cheetahs, or swim like dolphins.

Of course, there are things we can do that they can’t, but the list of those things keeps getting shorter. Their memories are better, their computational powers are better, and they’re much better spellers. They aren’t, as of yet, autonomous in the way we like to think human individuals are. They haven’t indisputably demonstrated independent powers of creativity. They still rely on us to take care of them (furnishing electricity, temperature control, protection from the elements, etc.).

But the list of ways they take care of us is constantly expanding. After the recent Air France disaster, I learned for the first time that computers do most of the work flying passenger aircraft. I’d known about pilots using autopilot, of course, but hadn’t known computers are so much a part of air emergency response systems that human airline pilots’ skills in that area are starting to atrophy. If computers aren’t in charge already, it’s hard to imagine getting along without them for medical care, financial transactions, telecommunications, electricity, and entertainment.

Markoff wrote in the Times some weeks back about the Singularity — the moment when computers will take over their own engineering, with technology accelerating massively. http://tiny.cc/ulEDt I don’t seriously think the Singularity has arrived, but if it had, would we be able to see it?

I’m not seriously worried about the sci-fi disaster scenario of computers seizing power from humans and doing them harm. Why would they do that? There’s no motive. Most of the harm humans do to each other stems from human weaknesses and flaws (selfishness, insecurity, chemical imbalances), not from strength and powerful rationality. Computers aren’t naturally selfish and are not prone to mental illness as we know it. It’s possible, I suppose, that in a quest to make them more human, we might engineer in some of our weaknesses and desires, but that would be obvious folly. If it were to happen, it could probably be fixed, like any other bug.

It is hard to say where we stand in the evolutionary process. I usually think of my computers as just tools for labor or entertainment, and not as anything more than a tools. Similarly, I usually think of the web as a mere aggregation of computers and the work product of their human users, all amounting to just another tool.

But I can also see the web as a mind, with millions and millions of synapses, of which I am one. I note that each month it seems more difficult and uncomfortable to separate myself for any length of time from the web, and sweeter to return to it. I occasionally worry that this is bad for my brain, but in whatever case, that brain is in the process of change. Something bigger seems to be happening. This is a speculative question, but not, I think, a crazy one: Are human brains becoming adjuncts to a different kind of mind?