'OpenAI cannot make AI safe on its own,' fired researchers warn
The ex-OpenAI researchers disputed the grounds of their firings and said the impacts will leave current workers afraid to speak up on safety risks.
TL;DR
- Three former OpenAI researchers claim they were abruptly fired last week.
- They released a letter titled 'OpenAI cannot make AI safe on its own,' stating their firings will have a 'chilling' effect on company culture.
- OpenAI stated an internal investigation found the employees violated company policies on handling sensitive information.
- The fired researchers dispute the reasons for their termination, stating they acted within their employment bounds and were communicating with third-party safety organizations.
- They recommended embedding third-party safety auditors, preserving monitorability of frontier models, and supporting open dialogue with the safety ecosystem.
- An OpenAI research leader agreed with the recommendations and stated that terminations were not for raising safety concerns.