… "In the test, the latest version of ChatGPT, model 5.2, & Claude were more likely to lead the user away from delusional thinking.
Etienne Brisson from the Human Line Project says this kind of research is limited & that they had heard from people who'd had mental health spirals on these latestmodels too.
In April, Elon Musk shared a post about delusions on ChatGPT, saying 'Major problem', but hasn't openly addressed the problem on Grok."