AI: Article: The Blind Spot in AI Safety

[Kim Locke]

https://www.techpolicy.press/the-blind-spot-in-ai-safety/

[Debra Turnbull]

"We cannot govern well what we have not learned to detect."

Yes, that descibes AI in a nutshell. Trying to catch something when no one knows what to look for...

[Chris Johnston]

@Kim-Locke @Debra-Turnbull

Yet another reason why allowing human-in-the-loop oversight to be tapered down, before the technology has matured sufficiently to provide a level of consistently and quantifiably reliable output is unacceptable.

Humans may be flawed as oversight, but better a flawed human than no oversight at all.