i think if killing pedestrians is a common occurrence for you then you should really stay off the roads.

how do we know the ChatGPT models haven’t crawled the publicly accessible breach forums where private data is known to leak? I imagine the crawler models would have some ‘follow webpage-attachments and then crawl’ function. surely they have crawled all sorts of leaked data online but also genuine question bc i haven’t done any previous research.