I wouldn't call it a token effort, they went to quite a bit of trouble to make GPT-4 safe. This is an active area of research too. At some point you need to pro

I wouldn't call it a token effort, they went to quite a bit of trouble to make GPT-4 safe. This is an active area of research too. At some point you need to prove GPT-4 would do something unsafe. If anyone did, they would improve their systems in response.