Robert Lemos at DARKReading just reported on a worrying trend. The title said it all, and the news is that more than 4% of employees have put sensitive corporate data into the large language model, raising concerns that its popularity may result in massive leaks of proprietary information. Yikes.