Romeo Dean proposes a verification framework to stop unauthorized AI training, estimating R&D at a few hundred million dollars · Digg