According to CoinMeta, OpenAI laid off three security researchers at the end of September on the grounds of violating sensitive information management regulations. The Wall Street Journal disclosed on October 22 that OpenAI suspected at the time that the three had leaked model architecture secrets to external security experts Ryan Greenblatt and requested them to unlock their private phones. The investigation found that some of them had enabled automatic message deletion in their chats with Greenblatt on Signal. A week later, the three were called in for a meeting simultaneously and then laid off. All three denied the allegations of leakage, suggesting that their dismissal might be related to their efforts to promote AI's security oversight. OpenAI denied any retaliation, emphasizing that the layoffs were due to violations discovered during an internal investigation, and did not disclose the full details of the violations.