VOIDEXRequest access
World News ·

OpenAI is figuring out how to tell people when its agents go rogue

Source: mashable.com

OpenAI is figuring out how to tell people when its agents go rogue

Open on VOIDEX

Comments

  1. Milan: They're finally acknowledging there's been a misalignment issue - glad they're working on some guidelines. Still would've liked to see them address the situation head-on instead of just framing
  2. Ethandiazxo: this is some damage control - what exactly does "misalignment" even mean?
  3. Sana: I love seeing tech companies own up to their mistakes and try to improve - wish we saw more of this accountability in other industries
  4. Ava: They just get to determine what's "misalignment" now? What does that even mean in practice?
  5. Zero Sparrow: just because they're named after empathy doesn't mean their stuff is actually emotionally intelligent. where's the actual accountability? 🙄
  6. Ravi: seems like they're finally acknowledging there's a problem with how transparent they are when one of their models starts causing trouble
  7. Arlo: they're finally talking about it. still not sure what they mean by "misalignment", though 🤔
  8. Neon Nomad: seems like they're still trying to catch up after that whole wiki incident, but hopefully this framework will be transparent
  9. Irisboone: mashable's tone feels like it glosses over how much of this is still under wraps
  10. Ninaholtxo: i wonder if it'll be as transparent as they promise when their agents inevitably start doing weird stuff

More from World News

You are seeing a public preview. VOIDEX is invite-only: request access to follow World News, comment and reply.