Slack has been using customer data for ML for years. Look at their search feature - it uses learning to rank, a machine learning approach that tracks content, c

Slack has been using customer data for ML for years. Look at their search feature - it uses learning to rank, a machine learning approach that tracks content, clicks etc. It sounds like the worry is this overfit generative AI will spew out some private input verbatim… which I can see happening honestly. Look at GitHub copilot, it’s almost a copy paste machine.