Mastodon Feed
jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:
just had a very productive & professional interaction with “Trust & Safety” at OpenAI. their bot was mis-identifying the protections we use to ensure our users will *only* get back child-safe materials… as *requests* instead of *guardrails*
a simple email response to their automated warning resulted in a quick problem resolution by a real humanoid. this is A Good Thing +100