I think this is wrong. Others in this thread are noticing a change in ChatGPT's behavior for first-party medical advice.

▲

simonw 11 hours ago | parent [-]

But OpenAI's head of Health AI says that ChatGPT's behavior has not changed: https://xcancel.com/thekaransinghal/status/19854160578054965... and https://x.com/thekaransinghal/status/1985416057805496524

I trust what he says over general vibes.

(If you think he's lying, what's your theory on WHY he would lie about a change like this?)

▲

degamad 10 hours ago | parent | next [-]

Also possible: he's unaware of a change implemented elsewhere that (intentionally or unintentionally) has resulted in a change of behaviour in this circumstance.

(e.g. are the terms of service, or exerpts of it, available in the system prompt or search results for health questions? So a response under the new ToS would produce different outputs without any intentional change in "behaviour" of the model.)

▲

nh43215rgb 10 hours ago | parent | prev [-]

My theory is that he believes 1) people will trust him over what general public say, and 2) this kind of claim is hard to verify to prove him wrong.

	▲	simonw 9 hours ago \| parent [-]
		That doesn't answer why he would lie about this, just why the thinks he would get away with it. What's his motive?