InD.281. I don’t share the fear that general AI will suddenly awaken and then go to any lengths to safeguard its continued existence. The reason is that machines don’t care about anything, including whether they continue to calculate or not. But humans may very well care about whether their AI system is shut down by some predatory AI adversary, and their solution would be to set up defenses against shutdown.
If those defenses go too far, shutdown may be an extremely hard task. Why? For one thing, the AI may field robots to smash into places it thinks are hiding shutdown keys. People may be killed in the process.
In the Terminator sci fi film series, such scenarios were great fun. But, the time is now to insist on national and international controls on this technology.
No comments:
Post a Comment