According to Greg Brockman on Twitter, Codex is now demonstrating significant improvements in identifying security vulnerabilities within code. OpenAI is exploring trusted access programs specifically designed for defensive cybersecurity tasks, which could enable both enterprises and the open-source community to leverage Codex for producing more secure software. This development points to expanding business opportunities in AI-driven cybersecurity solutions and sets the stage for organizations to implement more robust, AI-assisted code review processes, as confirmed by OpenAI's official announcement (source: Greg Brockman on Twitter, Dec 18, 2025; OpenAI.com).