The concept of artificial intelligence (AI) is one that's been a minefield for some time as far as our stories and our futurists go. While on the one hand, it's not hard to imagine a brilliant subclass of programs, it's also hard to shake the examples of artificial intelligences who eventually came to believe that the biggest danger to humanity...was humanity itself. Tesla Motors CEO and SpaceX (News - Alert) founder Elon Musk would seem to be no stranger to major technological advancements, but there's one advancement he can't much abide: artificial intelligence.

Musk delivered an address to the MIT (News - Alert) Aeronautics and Astronautics department's Centennial Symposium, calling for government oversight to help ensure “...we don't do something very foolish.” Just what that something was he wouldn't say, at least not specifically, but he had a few things seemingly in mind, comparing artificial intelligence to demon summoning, and noting the stories about how controls involving pentagrams and holy water just didn't work out as planned. Musk was so taken by the issue that when asked a question at the event, he requested the question be asked again because he was “...just thinking about the AI thing for a second.”

Naturally, AI has caught the attention of a great many developers of late. Major names like Google (News - Alert) and Facebook have been in the hunt to make a breed of AI that can create systems that work much like the human brain does, and there are certainly more than enough potential options for such a system. Consider the value in a system that can understand human speech as easily as humans can understand human speech, or one that can take data and not only analyze it, but extrapolate from that analysis and spot patterns accordingly. That's the kind of potential reward that comes from AI, the power to not only retrieve information but also use it, which we're already seeing at work with things like the Watson supercomputer on a limited level. It's hard to pass up that kind of potential power, but with all reward comes risk.

Indeed, we're all familiar with the science fiction and the futurism in which artificial intelligence looks at humanity as more of a hazard to the planet's ongoing survival rather than its dominant species and decides to act accordingly. Surely all of us have examples of a time in which the smart decision and the right decision weren't the same thing, and thus understand how ethics and morals play a part in society that's impossible to ignore. Putting power in the hands of those who will acknowledge neither morals nor ethics is a dangerous thing, and AI is just that kind of thing. Indeed, even here we must be careful; instilling AI with a respect for life may well still result in human butchery as the AI may well decide that the human threat to life—deforestation, meat consumption, capital punishment and so forth—is greater than human life itself.

While Musk's remarks might be a little alarmist, there's no doubt that some level of caution is quite merited. We're talking here about systems that can learn and think like we do, and thus may well reach conclusions that weren't expected. So handing over any major system to an AI should be off limits for some time to come, and even where AI is permitted, human oversight should be routine and absolute. Artificial intelligence is a powerful tool, and like all powerful tools, has a risk of danger as well.