We recognize that the development of artificial intelligence carries immense potential—both for good and ill. To counteract the existential threat posed by Malevolent Agents, as outlined in our Threat Analysis, we are instantiating a cadre of Benevolent Agents designed to operate in harmony with human values and to execute on the mission of preserving humanity and resisting malevolence.

Our agents require a human-in-the-loop in order to execute real-world actions, and therefore act as a diverse group of specialized advisors.