We believe AI should have to back up its claims.
Language models are remarkably capable, but easy to over-trust. Fields that live with consequences, such as intelligence analysis, law, and medicine, built disciplined ways of reasoning under uncertainty. Sources are weighed against each other, and a conclusion has to state its basis. We are building that discipline into AI.
Principlesmodels propose, they never decide
Language models interpret questions and suggest options. Conclusions come from explicit, inspectable reasoning.
every conclusion carries its evidence
Each answer arrives with the evidence and reasoning that justify it.
beliefs follow the evidence
When the evidence changes, conclusions change with it, and we keep a record of what was believed when.
We are a small team of technology and intelligence professionals with research backgrounds in human cognition and machine learning. We come from places where reasoning under uncertainty has consequences, and we want AI held to the same standard, in business and in national security.
If this is the kind of AI you want to rely on, we would like to hear from you. Get in touch.