Amid the escalating U.S.-China trade war, concerns over the security implications of competitive Chinese technology like Huawei’s 5G network, and unresolved negotiations with Beijing over the theft of intellectual property, another tech policy question persists in Washington, although it is somewhat overlooked. How should the United States manage exports of artificial intelligence technologies? It has widespread ramifications for global research, innovation and commerce—and no easy answer.
In November, the U.S. Commerce Department proposed a new rule on export controls for “emerging technologies that are essential to the national security of the United States.” Biotechnology, advanced computing technology and additive manufacturing—in other words, 3D printing—were all categories on the list to evaluate for prospective export controls. One of the biggest sections, however, pertained to artificial intelligence and machine learning technology. The controls would seek to limit the countries to which U.S.-incorporated firms could export artificial intelligence tools, such as computer vision algorithms, speech recognition systems, and audio and video manipulation technologies, which can be used to make fabricated but extremely realistic videos and photos using AI and machine learning, known as deepfakes.
The tech industry has pushed back adamantly on these proposals. “It is very broad,” Ed Black, the president of the Computer and Communications Industry Association, told Marketplace. “It is, in fact, potentially very sweeping and could impact, depending on how it evolves, a very wide number of companies, industries and a big part of the economy.” Black previously worked on export controls in the State Department and Commerce Department.