Longtime Slashdot reader Okian Warrior wrote: There has been a lot of discussion about AI regulation in recent news, with Sam Altman appearing on a Senate committee to plead for regulation. So far I’ve only seen calls for regulation, but no suggestions as to what regulation should be. Slashdot has a lot of experts in various fields (software, medicine, law, etc.), so maybe this discussion should start. And be aware that if we don’t create reasonable rules, Congress (mostly 80-year-old white men with conflicts of interest) will do it for us.
What are some good suggestions for AI regulation?
Let’s start. Humans (especially not AI systems) must be responsible for any treatment or diagnosis. If an AI suggests a diagnosis or treatment, it must have the buy-in of a human who believes the decision is correct and is as responsible as a non-AI doctor would be. AI should be a tool used by human decision making, not a replacement for human decision making. This avoids the problem of humans ignoring their responsibility, relying on software, and causing harm through negligence. Doctors can use AI to diagnose cancer (for example), but that’s the doctor’s diagnosis, not the AI’s.
What other suggestions do you have?
