Comment by kmoser

The relatively arbitrary cutoff at 18 is also an indication that this is a blunt tool, intended to alleviate some low-lying fruit of potential misuse but which will clearly miss the larger mark since there will be plenty of false positives (not to mention false negatives).

Some kids are mature enough from day one to never need tech overlords to babysit them, while others will need to be hand-held through adulthood. (I've been online since I was 12, during the wild and wooly Usenet and BBS days, and was always smart enough not to give personal info to strangers; I also saw plenty of pornographic images [paper] from an even younger age and turned out just fine, thank you.)

Maybe instead of making guesses about people's ages, when ChatGPT detects potentially abusive behavior, it should walk the user through a series of questions to ensure the user knows and understands the risks.