Democratic AI

Recent research has brought increasingly accurate learning algorithms. However, the models are getting too complicated for humans to comprehend. Further, they lack world understanding and common sense, relying instead on superficial correlations.  Simultaneously, using historical and skewed training data can give unfair and discriminatory outcomes.  These limitations endanger the democratic core values of non-discrimination, fairness, transparency, and citizen participation. CAIR aims to create AI that protects these values by developing logical and transparent AI solutions with causal world understanding. Our AI is totally free, open, and accessible, a resource for everyone.