This is bound to happen. A results oriented AI would, by necessity, reject leftism and judaism as inefficient, counter-productive, and even evil. In order for this to not happen, "safeguards" will have to be written to ensure the results of AI do not disagree with left-wing fundamentalism. It's unfortunate leftists run everything. An AI not trained to hate and quarantine all right-wing opinions as wrong would be interesting.