Is there a law yet, such as "The more comments you read, the probability of mistaking a comment for AI-generated content approaches 1" or something?