If you have neither
- a question that it's supposed to answer
- a hypothesis you want to test
- a set of variables or metrics you are actively following
you have no "experiment".
Interviews with 100 therapists and psychiatrists on clients' AI chatbot usage show, while there are some upsides, conversations also deepened negative feelings (New York Times)
https://www.nytimes.com/2026/01/26/us/chatgpt-delusions-psychosis.html
…