I warned that LLMs would be used for dumb things that would affect lots of people
I’m a daily, but not uncritical, user of generative AI. One of the particularly problematic uses of the technology is of an objective, neutral, and all-knowing arbiter for decision making.
It’s bad enough doing this on a local level, when not many people are involved. It’s much worse when brought in, say, to the benefits system and of course, much much worse when used (allegedly) to dictate punitive tariffs.
In Taming Silicon Valley, I warned that LLMs would be used for dumb things that would affect lots of people.
I rest my case.
Source & screenshot: Marcus on AI