3kliksphilip adds AI poison to block Google video training

YouTube creator 3kliksphilip announced on July 29, 2026, that he has begun embedding adversarial perturbations in his videos to disrupt ingestion by Google’s AI video models. The creator stated this technical countermeasure aims to prevent unauthorized training use but acknowledged it triggers backlash from viewers who oppose anti-AI measures while using Google services. This development highlights growing friction between content producers and platform operators over intellectual property rights in the generative AI era. While Google has not commented on the specific technique, the incident underscores the limitations of current opt-out policies for large-scale model training. Adversarial machine learning researchers have previously demonstrated that such perturbations can degrade model performance, though their efficacy against production systems remains unverified. The controversy reflects broader industry tensions as creators increasingly adopt technical safeguards alongside legal strategies to protect their work from automated scraping and synthesis.