Y
Hacker News
new
|
ask
|
show
|
jobs
by
Ephil012
1161 days ago
How do you prevent cases where the AI might say something harmful (e.g. suggesting the user kill themselves)?