Whistleblowers Say OpenAI Broke Promise to Rigorously Test AI for Danger Before Releasing
https://futurism.com/the-byte/openai-accusations-promise-test-ai
Whistleblowers Say OpenAI Broke Promise to Rigorously Test AI for Danger Before Releasing
https://futurism.com/the-byte/openai-accusations-promise-test-ai
3 Comments
Submission statement: how should we think about OpenAI’s promises given this information?
They say they prioritize safety, but it feels like every other week there’s some whistleblower saying that things look very different on the ground. And things could be even worse given all the disincentives to express concerns. Like that person who brought their concerns to the board and then was subsequently fired on trumped up charges. Or the NDA conditional on receiving *vested* equity, etc.
How do you think OpenAI is going to handle future security issues?
Who could have foreseen this sudden* but inevitable reveal**, that removing the super-alignment AI safety team wasn’t a sign they were doubling down*** on AI safety****!
No shit? I coulda have told you that after about 10 prompts.