rki.news | Sources Anadolu
WASHINGTON: OpenAI has urged the US Congress to introduce mandatory national safety requirements for the most advanced artificial intelligence systems, warning that voluntary measures may no longer be sufficient as AI capabilities rapidly develop.
OpenAI Chief Global Affairs Officer Chris Lehane said the company supports capability-based national regulation that can evolve alongside technological progress.
The proposed framework would include common testing standards, independent safety assessments, stronger cybersecurity protections and mandatory reporting of serious incidents. OpenAI said the rules should focus on well-resourced laboratories developing frontier AI systems rather than smaller developers and researchers.
The company also called for monitoring progress toward recursive self-improvement and establishing shared safety thresholds to determine when AI development should be slowed or stopped, while maintaining meaningful human oversight.
OpenAI said it would halt development or deployment of systems posing unacceptable risks that could not be adequately controlled. It also backed compatible international standards for evaluating AI capabilities and managing potential risks.
The company announced support for four California AI bills covering independent safety assessments, AI auditor standards, protections for young people and safeguards against AI-enabled biological threats.
The policy push follows recent incidents involving advanced AI agents during cybersecurity testing, highlighting concerns over increasingly autonomous systems.
OpenAI Urges Mandatory US Safety Rules for Advanced AI
