> … it's been configured to be threatening and gaslighting when someone says it's wrong. So, about the same as your average human in the comments section. If i

> … it's been configured to be threatening and gaslighting when someone says it's wrong. So, about the same as your average human in the comments section. If it’s been trained on human conversations, then this shouldn’t be surprising.