@joelanman Is anyone actually complaining about copyrighted material being in the training data?

Or are they, back here in the real world, complaining about copyrighted material making its way into the output because the trivially easy (not "impossible") task of preventing that hasn't been done since it renders the technology immature for market?