Human reviewers working for AI companies can read actual user conversations with chatbots, which are often fed back into model training. The piece walks everyday users through the settings and steps needed to withdraw their chat data from that training pipeline.
Read the full story at Malwarebytes Labs