Hacker News new | past | comments | ask | show | jobs | submit login

That's still putting a lot of trust into the AI that it's not going to overstep in some way that's problematic to you, and given the lawyers that have to sift through the content you create may not be very happy about it, they might be more than willing to jump on any excuse to cause you problems based on what you submitted.

As an example, just search about ChatGPT being "confidently wrong" or "confidently incorrect". It's also possible that in restating things ChatGPT may be "colloquially accurate" but "legally incorrect" in how it restates things, given it's trained on many sources of natural language, and I think most people would choose to communicate with a court and lawyers (especially ones they assume antagonistic) with a higher level of caution than colloquialisms as they would when speaking informally to others.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: