Is Google AI alive? (13 Jun, 2022)

[aardvark_admin]

This column is archived at: https://aardvark.co.nz/daily/2022/0613.shtml

Are we really reaching the point where our AI systems are becoming sentient?

Or is this just another shock/scare story based on hype rather than fact?

More importantly, do AI systems need to become sentient to actually pose a threat?

Are Google doing enough to ensure that their AI is properly trained?

How long before we start finding that we've created a huge problem that may be very difficult to unwind?

[GSVNoFixedAbode]

I'd put my money on a very sophisticated language processing system rather than sentience, although you need to be able to clearly define sentience itself before using it as a yardstick. That said, we still need an independent Big Red Button wired in to cut power in case of rogue AI, sentient or following a badly-defined ruleset.

[Perry]

"Google doesn't like spending time or money. It simply wants to collect the ad revenues."

What spending that collected money on is far more important than the collection.