Federal appeals court ruled open-source programmers cannot proceed with claims that Copilot failed to provide proper code attribution, eliminating a major copyright threat to AI training practices. This marks a significant shift in copyright liability risk for model developers and establishes precedent that attribution disputes alone don't sustain claims against AI coding tools.

Australia's proposed government copyright rules heavily favor AI companies over content creators, with former competition regulator Rod Sims warning creators face payment elimination. This signals a major policy shift that reduces compliance friction for AI training infrastructure in a key jurisdiction while creating international regulatory arbitrage.

As of August 2, 2026, the EU AI Act entered active enforcement with Article 50 transparency requirements now binding, and December 2 deadlines for prohibited practices and Annex III high-risk system compliance. Practitioners building in Europe must now operationalize watermarking, metadata documentation, and third-party audits or face enforcement action from the AI Office and national authorities.

On September 7, China's Supreme People's Court issued judicial guidance with 24 provisions directing courts to apply existing laws to AI-related disputes, establishing the first national precedent framework. This clarifies liability and contract enforcement for AI systems in China's jurisdiction, reducing legal uncertainty for builders operating in or training on Chinese data.