OpenAI reasoning monitoring letter

The Wall Street Journal reported on October 7 that it reviewed a letter from three recently terminated OpenAI researchers to the company’s board and safety committees. The former employees asked OpenAI to preserve the ability to monitor the reasoning of increasingly sophisticated models and to work with outside safety auditors. Their request expresses concern about a possible future loss of visibility; the report does not show that monitoring has already been lost or that OpenAI has adopted the proposed measures.