Trump administration filed a 20-page brief in Manhattan federal court arguing that training LLMs on copyrighted material qualifies as fair use and that restricting AI training threatens national security and economic competition. This is the first federal government position on unlicensed AI training data and signals major policy shift that could reshape copyright litigation risk for all LLM developers.

DOJ argues that AI training is 'extraordinarily transformative' and that LLM outputs do not directly compete with source copyrighted works, meeting fair use standards. Developers now face reduced legal exposure for training data practices if courts adopt this framework, though international and state-level litigation continues.

European Commission designated ChatGPT as a Very Large Online Search Engine (VLOSE) under the Digital Services Act, with compliance deadline December 2026 and penalties up to 6% of global annual turnover for non-compliance. This creates parallel compliance burden for OpenAI and sets precedent for categorizing generative AI systems under DSA rather than AI Act alone.

Trump administration framed AI training restrictions as threats to American competitiveness and cited national security as rationale for backing OpenAI's fair use defense against media lawsuits. This positions copyright liability as a strategic competitive issue, signaling developers they can expect favorable US policy environment while EU tightens oversight.