I keep seeing more disturbing and funny responses by various AI tools. They betray some kind of hypersensitive constraints programmed in to it, apparently by people who are terrified of it producing anything offensive to anyone. Which as a result, is offending lots of people.
While I, too, find these AI constraints odd and creepy, companies misread their markets and make strange decisions all the time. The solution is either for enough people stop patronizing them or criticize them until they better align to the market, or for people to create competitors to better serve other segments.
Criticism is fine and sometimes spurs broader conversations about implicit and explicit cultural values, but at the end of the day, companies ought to be able to make stupid things.
More dangerous than companies that make stupid things is the notion that companies should be treated as “public utilities”. AI bots, just like heads of lettuce or tennis shoes, are created by companies and given or sold to voluntary users. They don’t have to not be stupid. They don’t have to be reasonable at all.
We can point out the bizarre behavior of these bots, and it is of some benefit to do so, but the best solution is to build a better one.