- Watch out for the em dashes, “—” in LLM outputs. Some LLMs, including ChatGPT, tend to use this special version of dash in its writing. If using LLM to write, you might want to replace the “—” with “-”. People may see the long dash and assume AI wrote the entire thing if they know about this pattern.
- There are tools such as https://gptzero.me/ that can detect how like something is written by AI. Could be useful both as the writer and the reader.
(Going to add more when I see them.)