OpenAI has introduced GPT-5.4-Cyber, an advanced large language model tailored for cybersecurity applications, alongside an expansion of its Trusted Access for Cyber (TAC) program. Launched on April 14, the updated TAC initiative aims to automate identity verification and streamline cybersecurity processes while collaborating with select organizations. After months of iterative improvements, OpenAI has opted for a staggered rollout of GPT-5.4-Cyber to assess the model’s benefits and risks in real-world scenarios. The TAC program now has multiple tiers, offering users enhanced capabilities for cybersecurity workflows while ensuring stronger verification protocols to prevent misuse. As AI can be dual-use, OpenAI emphasizes the importance of secure software development. The company envisions a dynamic ecosystem that identifies and mitigates security vulnerabilities continuously, integrating AI-driven tools into coding workflows for real-time risk reduction. This shift aims to transform how cybersecurity challenges are addressed in the software development lifecycle.
Source link
