From the source
OpenAI published a blueprint for governments and industry on putting advanced AI into cyber defenders' hands, arguing there is a limited "defender's window" before offensive cyber capabilities spread widely.
- The company says its own experience with misalignment and security incidents involving its models shows the need for stronger safeguards and accountability across the industry.
- The blueprint sets out proposals in three areas: clearer responsibilities for securing frontier models and agents, including detecting dangerous behavior and maintaining reliable shutdown controls; equipping defenders with tools, training and technical support while fixing longstanding cybersecurity weaknesses; and strengthening government evaluations, independent technical assessments, consistent incident reporting and coordinated response.
- OpenAI says AI labs have a responsibility to address the risks their systems create, while governments need the expertise, resources and authority to protect the public as capabilities advance.
- It encourages policymakers to act now and to tighten requirements as the technology evolves.
