According to CoinMeta, OpenAI has begun a large-scale overhaul of its internal security. Thibault Sottiaux ( Tibo ), who is in charge of product development, took personal charge of the effort, mobilizing about a quarter of the engineers to specifically fix security vulnerabilities in the internal systems. On November 2nd, OpenAI announced a partnership with a cybersecurity company, Cathedral, to launch a new round of security tests and increased the bounty for serious vulnerabilities to more than four times the original amount. Sources interviewed by Reuters revealed that long before the ChatGPT announcement, employees had repeatedly raised concerns about internal security issues with management, arguing that the company's management of code and research materials was too lax and that security checks were not strict enough. In 2023, a hacker managed to intrude into OpenAI's internal Slack and steal technical details of the GPT-4 model. As AI technology has become a strategic resource contested by various countries, OpenAI has had to re-evaluate its internal security and invest more resources in vulnerability fixing and protection efforts.