Comment by mikkupikku

As well as the points already raised by others, I'd like to make the point that we should be encouraging people to prompt LLMs themselves rather than just accepting the outputs of others. As a social norm, this will make society more robust to misinformation and deception, as it will result in fewer people trusting outputs without knowing how the LLM was actually prompted.

This probably doesn't really matter in this context, but I think it's a general best practice worth reinforcing whenever possible.