[RandomGerm4n]

I think the better solution would be to treat every answer as if it were written by a human and then hold the person accountable for any mistakes. I’ve often seen in academic contexts that people have published texts with completely fictitious citations. In such cases one should confront the person as if they had done it on purpose. No one can accidentally invent entire books and authors. The same goes for software projects, where some people submit code that clearly comes from an LLM and hasn't been properly reviewed beforehand. If there are security vulnerabilities in that code, you can accuse the person of having done it on purpose. Anyone who submits AI code without labeling it as such is affirming that they understand the code and have thoroughly reviewed it.