Yep, kill any content an ML model has predicted of high risk or unprofitable. Classic example of a regulation second order effects resulting in the opposite outcome of what was intended.