Can companies police the biases found in artificial intelligence?