I did try asking one of the PIs whether they were doing it to demonstrate that it doesn't work in order to get paper authors to provide their real data but didn't get a reply.

Will be going to a conference at the end of the month where there will be several presentations on the use of LLMs for this.