Responsible AI
Our software supports human judgment. It does not replace it.
Decisions that involve the use of force are always made by humans.
Outputs are traceable to their sources, so people can review what the software used and why.
Customers and use
We provide our software to legitimate government institutions of the United States and its allies, and to authorized customers.
We do not provide it to sanctioned parties or to adversarial states.
We may decline or end work that is inconsistent with these principles.