While OpenAI acknowledges that some people have an unhealthy relationship with its chatbot ChatGPT, the company doesn’t think those problems are particularly widespread. (It also thinks its product will eventually be good enough to deal with those dicey situations.)
CEO Sam Altman, in a rare on-the record dinner with journalists, told The Verge that “way under 1 percent” of users have problematic relationships with it.
Even if that’s true, let’s hypothetically say 0.5% of users have acutely negative outcomes from their conversations with ChatGPT. For a product that now has 700 million weekly users, that’s still 3.5 million people.