Attribution Decay and Mass Scraping Challenge AI Copyright Claims

MIT research indicates that removing one artist from training data barely changes model outputs, making individual copyright attribution difficult to prove. Reported scraping of roughly 12 million artworks from Cara reinforces unresolved problems around consent records, portfolio exposure, platform accountability, and effective anti-scraping defenses.