Microsoft chief economist Michael Schwarz told attendees of the World Economic Forum Growth Summit today that "we shouldn't regulate AI until we see some meaningful harm that is actually happening, not imaginary scenarios." https://arstechnica.com/tech-policy...ys-microsoft-exec/?comments=1&comments-page=1 Here is some costly harm to consider.. The Boeing 737 MAX passenger airliner was grounded worldwide between March 2019 and December 2020 after 346 people died in two crashes. The Department of Justice announced that Boeing would pay a $2.5 billion settlement. Why did they crash? The MCAS system took control away from the pilots. MCAS was designed “to address potentially unacceptable nose-up pitching moment at high angles of attack at high airspeeds.” The failure analysis didn’t appear to consider the possibility that MCAS could trigger repeatedly, as it did on both accident flights. MCAS was not included in the pilot manual at Boeings' request and Boeing won the FAA’s approval (the regulator) by recommending pilots get an hour of training on an iPad in regards to the differences between the MAX and the previous 737 generation - MCAS was not mentioned in that training. No simulator or in-flight training on MCAS (what triggered it, how to appropriately respond or how to shut it off) ever took place. Regulation would have saved 346 lives.
I will NEVER get on one of these. That said I would not ride in a self driving car either. I just hope to never get hit by either one. Maybe some day the tech will be good enough. Today is not that day.
Nuclear non-proliferation wasn't much of a success. Bioweapons and chemical warfare are still around. Practitioners of gain-of-function research are still funded. About 30 countries, according to Google's Bard, haven't signed the Anti-Personnel Mine Ban Convention. Why should people believe that AI can be meaningfully regulated?
Just because we can't craft the perfect regulation doesn't mean that we shouldn't craft ANY regulation. The purpose of rules and regulations is not to completely eliminate a problem, but to reduce the impact of that problem.
To reduce the possibility of reductio ad absurdum, I wrote "meaningfully regulated" rather than just "regulated".
