HabitableZone

Flame » in reply to Would you trust yourself to build an ethical AI?

Of course not.

Wanting to benefit extravagantly from your work does not automatically make you a selfish criminal. It just makes it easier for you to come up with excuses to be one.

The whole point of building an AI would be to create an autonomous, independent intelligence. And that means you really have no way of knowing how it will turn out, either ethically or in any other respect.

Building an AI would be like having a child. No matter how much positive influence you could exert on its development, or how conscientiously you tried to do the right thing, you would have no way of predicting how it will turn out. Remember Mary Shelley's "Frankenstein". Evil can just as likely be due to ignorance as malice.

So tell me, what kind of AI is more likely to turn out (unexpectedly) to be a toxic monster? One designed to navigate an interstellar spacecraft on a mission lasting centuries, or operate an automated dredging barge in the Amazon estuary. One built to manage an insurance company, run a media empire, or command a carrier battle group?

And remember, an AI would not only be able to act on its own, it would be able to EVOLVE on its own, independent of its designer's most benign motives.

Log in or register to post.

The whole thread (42 posts)

Related discussions