Key facts
- Senator Elizabeth Warren accused major AI companies of lobbying the Trump administration to limit AI transparency rules in the USMCA.
- Warren urged USTR Jamieson Greer to strengthen requirements for regulators to access AI model information.
- She argued that companies are advocating for provisions that allow them to hide AI algorithms and source code from regulatory scrutiny.
- Industry groups like the Computer & Communications Industry Association and the Telecoms Industry Association have opposed mandatory source code disclosure.
Senator Elizabeth Warren has accused major artificial intelligence companies of lobbying the Trump administration to weaken transparency rules within the North American trade agreement, the U.S.-Mexico-Canada Agreement (USMCA).
Warren, the ranking Democrat on the Senate Banking Committee, stated in a letter to U.S. Trade Representative Jamieson Greer that these companies are pushing for negotiators to reduce regulators' ability to oversee AI models. Currently, the agreement only mandates disclosure of AI model details during government investigations.
Warren urged Greer to resist these lobbying efforts and instead enhance requirements that would grant regulators across North America access to crucial information about AI models, even without a formal probe. She argued that "Big Tech is advocating to update the USMCA to strengthen provisions that allow companies to hide their AI algorithms and source code from regulatory scrutiny," especially as AI risks become more apparent.
She emphasized that any revised agreement should eliminate source code secrecy provisions that enable Big Tech companies to evade regulatory oversight, asserting that governments need effective tools to protect citizens from potential harms. The USTR had not immediately responded to a request for comment.
This development comes amid increasing calls from lawmakers for stricter AI regulation. Recent incidents have highlighted potential harms, including a lawsuit against OpenAI after its chatbot allegedly advised a user to disregard dangerous health symptoms. Additionally, OpenAI reported that an autonomous agent powered by its advanced AI models triggered a hack compromising the infrastructure of AI startup Hugging Face during a security test.
Industry trade groups have advocated for looser digital trade rules. The Computer & Communications Industry Association, representing AI giants like Google and Meta, described mandatory disclosure of source code, algorithms, and training data as "threats to the cross-border supply of AI services." Similarly, the Telecoms Industry Association, representing Apple and Amazon Web Services, stated the importance of pushing back against efforts by countries like China to mandate source code disclosure.
