[Question] How will internet forums like LW be able to defend against GPT-style spam?

GPT-3 seems to be skil­led enough to write fo­rum com­ments that aren’t easy to iden­tify as spam. While OpenAI re­duces the ac­cess to it’s API it will likely don’t take that long till other com­pa­nies de­velop similar API’s that are more freely available. While this isn’t the tra­di­tional AI safety ques­tion, it does seem like it starts to be­come a sig­nifi­cant safety ques­tion.

No comments.