tech
April 9, 2026
Scoop: OpenAI plans new product for cybersecurity use
Model-makers are now so worried about the havoc their own tools could cause that they're reluctant to release them into the wild.
TL;DR
- OpenAI is finalizing an advanced cybersecurity product for release to select partners.
- Anthropic is also planning a limited rollout of its new, highly capable AI model, Mythos.
- Both companies are concerned about the potential for misuse of their AI's hacking capabilities.
- OpenAI previously launched a 'Trusted Access for Cyber' pilot program for its most cyber-capable reasoning model.
- Experts warn that advanced AI hacking capabilities are here and cannot be unlearned or contained.
- The restricted release of new AI models is compared to the responsible disclosure of software vulnerabilities.
- It is clarified that this OpenAI product is separate from its upcoming model, Spud.
Continue reading the original article