Artificial intelligence policy, discussions around AI risk and safety, and intellectual property issues have collided ahead of the scheduled September 24 meeting between U.S. President Donald Trump and Chinese President Xi Jinping.
Several leaders and researchers at major AI companies recently raised concerns about the speed at which increasingly capable AI systems are being developed. Public discussion has included proposals for stronger safety testing, common technical standards, greater disclosure of serious incidents and, in some cases, slowing development of the most advanced systems.
Despite these cautions, the U.S. government has generally continued to emphasize maintaining American competitiveness in artificial intelligence while considering approaches to AI safety. On September 22, President Trump stated that the Department of Justice could intervene with AI companies, if necessary, while reiterating his opposition to broad new AI regulation.
International interest in AI oversight also is increasing. On September 22, U.N. Secretary-General António Guterres called for international oversight of advanced AI systems. Reuters also reported that Chinese AI developer DeepSeek is expected to brief the U.N. Security Council this week, together with representatives associated with U.S.-based OpenAI and Anthropic, regarding risks presented by advanced artificial intelligence.
While industry and governmental leaders continue the debate over AI, intellectual property issues surrounding AI model development remain unresolved. One increasingly important matter is “model distillation.” Distillation is a commonly used AI training technique in which outputs from a more capable model are used to help train another model. While the technique itself is not inherently improper, questions arise when a proprietary model is accessed or used without authorization to develop a competing model. Anthropic recently reported that it had identified and disrupted what it characterized as unauthorized distillation campaigns involving several China-based AI laboratories. Anthropic stated that those activities involved fraudulent accounts, proxy services and large numbers of interactions with its Claude models. The identified companies and Chinese officials have disputed or rejected the U.S. allegations concerning improper model extraction.
Distillation itself raises potentially significant questions under trade secret law, contract law, copyright law, computer-access statutes, and other existing legal regimes. Current U.S. and international laws do not yet provide a single legal framework specifically addressing unauthorized extraction of AI model capabilities.
The United States and China also have begun discussing AI safety directly, with officials from both sides agreeing to continue AI safety discussions and establish mechanisms for communicating about certain AI incidents. Artificial intelligence, advanced semiconductor technology, and related security issues are expected to be among the topics addressed in the September 24 Trump-Xi meeting.
For now, AI intellectual property protection, model distillation, AI safety, and the developing legal framework remain in flux. Lando & Anastasi will continue monitoring developments in these areas and will provide further updates as circumstances evolve.
This IP Advisory was prepared by Lando & Anastasi, LLP. The information provided in this Advisory does not, and is not intended to, constitute legal advice; instead, all information, content, and materials are for general informational purposes only. Readers should contact an attorney to obtain legal advice with respect to any particular legal matter.
© 2026 Lando & Anastasi, LLP
