Please no, LLMs have been shown to introduce biases from training data into the generated "summaries". LLMs should go nowhere public policy.