@ClementDelangue: Agree with @gdb that we need to arm cyber defenders much more than they are now, both with APIs and open models. The re…
Summary
Clement Delangue agrees with Greg Brockman that cyber defenders must be equipped with AI tools like APIs and open models to address the asymmetry in capabilities between attackers and defenders.
View Cached Full Text
Cached at: 08/21/26, 05:04 AM
Agree with @gdb that we need to arm cyber defenders much more than they are now, both with APIs and open models. The real cybersecurity risk of AI is asymmetry of power, capabilities and ressources between attackers and defenders!
Greg Brockman (@gdb): defenders can see the future, and have a narrow window to uplevel their cybersecurity practices now.
key is to uplevel fundamentals and apply the best AI tools.
what we’re doing at OpenAI, and where other organizations can start:
Similar Articles
@ClementDelangue: The cybersecurity debate on open-source AI is backwards. Open models aren't the risk, they're the defense! Attackers ca…
Clement Delangue argues that open-source AI models are not a cybersecurity risk but a defense, as attackers can already jailbreak closed systems while defenders need transparency to secure AI.
Strengthening cyber resilience as AI capabilities advance
OpenAI publishes a comprehensive framework for managing cyber capabilities in AI models, noting significant improvements in CTF performance from GPT-5 (27%) to GPT-5.1-Codex-Max (76%), and outlining defense-in-depth safeguards to ensure advanced models primarily benefit defenders while limiting offensive misuse.
@OpenAI: Advanced capabilities require strong safeguards. That’s why access is limited to approved defenders, with additional co…
OpenAI is expanding Daybreak with two access tiers (Blue and Red) and introducing GPT-5.6-Cyber, a purpose-trained cybersecurity model that significantly reduces refusals for authorized defensive security work.
@elonmusk: Extremely important point
Jensen Huang emphasizes that defenders need a frontier AI ecosystem combining open and closed models, citing a Hugging Face incident where an open-weight model helped contain an intrusion that closed AI blocked.
@sama: please consider using our models to help defend your systems
Eric Wallace announces the release of GPT-5.6-Cyber, a model focused on cybersecurity tasks like exploit development, and urges Sam Altman to consider using it for defensive purposes.