On Tuesday, May 14, officials from the United States and China held their first formal bilateral meeting in Geneva to discuss the “technical risks” of artificial intelligence. However, the meeting revealed significant disagreements between the two sides on the development, use, and establishment of common standards for managing artificial intelligence, indicating that a consensus on safety standards for AI technology will not be reached in the short term.
The White House described the meeting as an information-sharing session, which included details of U.S. domestic policies regarding the risks and security of artificial intelligence, as well as various countries’ approaches to managing this technology.
According to Chinese state media reports on Wednesday, the Chinese government described the meeting as “in-depth, professional, and constructive,” but also emphasized a firm stance regarding U.S. restrictions and pressures on China in the field of artificial intelligence.
In a brief statement released after the meeting, the White House expressed concerns about the potential misuse of artificial intelligence, including by China, and its implications such as interference in elections, the creation of bioweapons, and launching cyber attacks.
A senior official from the Federal Bureau of Investigation (FBI) expressed concerns last Thursday that foreign adversaries, especially China, Russia, and Iran, may utilize artificial intelligence to interfere in the upcoming U.S. elections in November and spread false information.
Given the prevalence of deepfakes and robocalls, coupled with the emergence of such technologies in the political arena, the threats posed are not merely theoretical.
ABC News reported an incident earlier this year in Slovakia where a deeply faked audio clip resembling a party leader’s voice discussing raising beer prices and manipulating votes was widely circulated on social media platforms just days before parliamentary elections.
Moreover, in April, Microsoft disclosed that “China is using fake social media accounts to conduct opinion polls on voters, identify their areas of disagreement, sow discord, and potentially influence the outcome of the U.S. presidential election in its favor. China has also expanded its use of artificially generated content to advance its global goals.”
The U.S.-China Economic and Security Review Commission warned that China is leveraging advancements in commercial artificial intelligence to enhance its military capabilities in preparation for a potential conflict, particularly regarding Taiwan.
“Politico” quoted Sam Bresnick, a researcher at Georgetown University’s Center for Security and Emerging Technology and an expert in Chinese technology policy, who stated that compared to the U.S., China is unwilling to address the risks associated with artificial intelligence, which is a major concern.
Bresnick noted that the Biden administration is apprehensive about China’s assessments and testing agreements on artificial intelligence not meeting the necessary standards, suggesting that China may be more inclined to use untrustworthy or inadequately tested AI systems.
The potential for chaos triggered by artificial intelligence is a tangible concern. Jacob Stokes, a senior researcher at the Center for a New American Security focusing on China issues, warned that civil AI systems could inadvertently cause “disastrous events… shutting down power grids or altering satellite orbits, leading to collisions or disrupting critical services.”
The Biden administration has taken steps to monitor the actions of foreign adversaries, including China, in the realm of artificial intelligence.
In a statement in January, the White House urged U.S. cloud computing companies to promptly inform the government if they detect foreign customers using their services to train powerful AI models that could be used for cyber attacks.
The U.S. Congress is calling for a stronger response from the government to address the potential threats posed by China’s use of artificial intelligence.
On Monday, May 13, the co-chairs of the House Intelligence Committee on U.S.-China Competition, John Moolenaar and Raja Krishnamoorthi, introduced a bipartisan bill titled the “Enhancing Foreign Countries Oversight of Sanctions Enforcement Act.” This legislation aims to prevent the Chinese Communist Party, its military, and affiliated companies from gaining access to AI systems that could be used for future cyber attacks against the U.S.
Moolenaar stated that “AI is a foundational technology of the future that could reshape our economy, way of life, and national security. This legislation will protect the American people and help prevent our country’s innovations from falling into the hands of adversaries.”
The Senate’s bipartisan AI Working Group released a “AI Policy Roadmap” on Wednesday, urging the U.S. government to “avoid creating policy vacuums that China and Russia could exploit to gain more influence and control.”