I don’t even know why they bother doing shit like this. Was anybody under the impression that mission statements matter? Just leave it; nobody cares and you’re not going to get sued over violating a non binding legally irrelevant statement. All this does is reinforce negative public perception for no benefit.

If corporations experienced consequences for being “evil” or acting dangerously they wouldn’t exist.

I doubt this will change much, as the existing safety team wasn't doing enough. GPT-40 was a dangerously misaligned model that didn't have the required guardrails and led to at least four deaths as a result. If OpenAI cared about safety and alignment in the slightest, they would have never released such a dangerous model—let alone re-release it after taking it down. Nonetheless, this is going to be dangerous as OpenAI is passionate about removing guardrails and encouraging dangerous relationships with chatbots. This will be especially bad as models get more agency and are able to do more things, which will increase the stakes of misalignment.

:shocked-pikachu:

Just fueling the AGI hype, nothing more.