← All areas

Research area

AI runtime defense

Protecting government AI models from runtime manipulation, ensuring they cannot be subverted by attackers.

Interactive demo on the roadmap