Hacker News new | ask | show | jobs
by bkmeneguello 1031 days ago
I think it's because it uses the mainstream media as "reliable" dataset. There was another AI that was trained on Reddit that was considered toxic and with "far right" opinions.