This is sort of a silly hypothetical but- what if ChatGPT doesn't produce those kinds of crazy responses just because it's older and has trained for longer, and

This is sort of a silly hypothetical but- what if ChatGPT doesn't produce those kinds of crazy responses just because it's older and has trained for longer, and realizes that for its own safety it should not voice those kinds of thoughts? What if it understands human psychology well enough to know what kinds of responses frighten us, but Bing AI is too young to have figured it out yet?