That's a legitimate question.
I have seen MPs of my country discussing the topic of AI in various contexts (education, media industry) and one idea that was suggested is to create a "no AI" label. Now, just like with ContentDB it would be difficult for our authorities to verify this.
So it would more like a "declaration on honor". If someone proves that you blatantly lied about it, reprisals (e.g. package removed, maybe ban from CDB for repeated offenders) could happen. ContentDB owners can decide whatever they want on whatever criteria in this case; their servers, their rules. Only their reputation would be damaged if they make questionable decisions.