Presenter lays out two principal AI risks: misuse and loss of control

The presenter told the council that AI's rapid capability growth creates two main categories of risk. "The 1st is misuse of AI models," he said, citing misuse by bioterrorists as an example; "The 2nd risk is loss of control, model capabilities accelerating beyond developers' ability to control them."

He argued these risks are global in scope and that managing them requires state-level attention beyond economic or information-technology ministries. The presenter used the risk framing to justify proposing international standards, verification regimes, and an incident-notification system for AI harms that could rise to global-security significance.