Mitigation must be reasonable, proportionate, effective, tailored to the specific Article 34 risks, and assessed with particular consideration for fundamental-rights impacts. The evidence should explain why the measure fits the risk and why a less intrusive or more targeted measure would not achieve the same result.
Article 35 examples include adapting service design and interfaces, terms and enforcement, content moderation processes and resources, algorithmic and recommender systems, advertising systems, internal processes and supervision, cooperation with trusted flaggers and dispute-settlement decisions, codes of conduct or crisis protocols, user awareness and interface information, child-protection measures, and markings or reporting functions for generated or manipulated media.