I don't think the idea is that the speech in the chat is inherently illegal; it's that it could be used as evidence of illegal activity. Using that example - if someone in the chat asks about plate XYZ at 10AM, and if a phone linked to "Bob" posts to the group chat at 10:04 AM that license plate XYZ is used by ICE, and the internal logs show that Bob queried the ICE database about plate XYZ at 10:02 AM, and no one else queried that license plate in the past month, that is pretty good evidence that Bob violated the CFAA.