Hopefully they review it very carefully and not well after the event so the memory is fresh.

Otherwise, you may forget important details and just think it was correct if you need to reference it later. We know how influential someone's memory can be.

There is something to be said about needing to write this stuff down fully to commit it to memory.

I agree that this is likely going to end badly.

You think AI chatbots are bad with facts, you should see the cops!

Honestly I don't think this will be any worse than police filing hand-written reports rife with their own biases and material omissions and sometimes even outright, intentional lies.

Using LLMs to write a first draft based on your notes, rephrase and reword existing text is completely different from using an LLM to recall facts in its training data.

I see this take a lot but I am not sure its entirely true. If we were to blindly ask questions based on a models world knowledge, yes there will be false statements. When asking questions or summarizing data across a provided dataset, I have not seen this to be a problem. Now I think there could be an interesting question on how its summarizing, there may be inherent biases but I don't believe the original issue is in fact a problem.