Clearly tagging AI generated output to avoid unintentional misinformation

The topic of hallucinations came up in one of the lectures and I’m wondering if there are any standards or best practices right now to tag or label AI generated output as AI generated so that we don’t unintentionally cause misinformation. I’ve started using a string " :robot: Generated by an AI bot" for the prototype I’m building but wanted to check if there were any existing standards or best practices.

I don’t think there is a standard yet. It’s an area of rapid development and change.

1 Like

That’s a good question and I think something with a lot of opportunities for finding a clever solution to. Some groups are working to implement LLM’s with hidden signatures built into the text content itself. Though there isn’t a standard yet, I like your idea of adding a string to indicate it’s a bot! (:slight_smile: