Gil Gekker, Rachel Steratore, Everett Smith, Asher Brass-Gershovich, Varun Gandhi, Nicole Nichols, Vijay Bolina, Buck Shlegeris, Lisa Einstein, Dan Lahav, Omer Nevo, Sella Nevo
Author
Gil Gekker, Rachel Steratore, Everett Smith, Asher Brass-Gershovich, Varun Gandhi, Nicole Nichols, Vijay Bolina, Buck Shlegeris, Lisa Einstein, Dan Lahav, Omer Nevo, Sella Nevo
Date
Key takeaways · AI-distilled
The agenda splits AI security into four buckets: policy foundations, public-private coordination and institutions, technical security engineering and assurance, and governing agentic AI under adversarial pressure — agents treated as a problem apart from model safety.
Priorities were ranked on two axes at once, importance and cost-effectiveness, so the list surfaces cheap high-leverage work rather than only the biggest problems. The inputs were structured interviews across industry, government, and civil society, then a cross-sector workshop.
Each priority is written up the same way — define the problem, assess the current landscape, name actionable projects — which is a deliberate move to make the field enterable, so newcomers can pick up concrete work instead of only established practitioners.
Why it matters
It maps where the gap between AI adoption and AI security readiness is widest, which is the agenda teams securing agents will be measured against.
Key quotes
“As AI systems are rapidly integrated into critical economic, governmental, and national security functions, the gap between AI adoption and AI security readiness continues to widen.”
“Participants identified and ranked the highest-importance and most cost-effective areas where progress could strengthen AI security - from protecting frontier AI systems and their underlying infrastructure to improving cybersecurity practices as AI reshapes the threat landscape.”
