This fake image of a man lurking outside a polling place with a gun was created by artificial intelligence tool ChatGPT Plus
Ever since the explosive development of ChatGPT, there has been widespread concern about potential disruptive implications of generative AI technologies on the political process and national security. As the 2024 U.S. election turns out to be the first “AI election” in history, how to mitigate the risks of unregulated AI and ensure election security has become a rare bipartisan issue in the United States.
As AI technologies pose opportunities and challenges with global dimensions, how to better unleash their potential and stem their harm have become prominent areas in which the interests of China and the U.S. converge. With the shared concern of AI disruptions, a sustained effort to make the best of a bilateral AI dialogue would be beneficial to address challenges.
First AI elections in U.S.
Heading into the 2024 U.S. elections, the potential use and misuse of generative AI has sparked heated debate. Some point to the great promise of AI in facilitating the election process, but more are concerned about the unprecedented perils of AI-generated deepfakes and disinformation in harming U.S. election security.
As the election proceeds, AI disruptions have already become a reality. Both political parties have embraced AI in their campaigns for messaging and mobilization, but there are also deliberate efforts to undermine opponents, mislead voters and even suppress voting. This is vividly shown in the AI-generated images of former president Donald Trump being arrested, the robocalls impersonating President Joe Biden calling for voters not to vote at the New Hampshire Democratic primary, the Republican National Committee’s AI-generated ad attacking Biden and more recently the voice-clone video mimicking Vice President Kamala Harris affirming Biden’s alleged “senility” and calling herself a “diversity hire.”
These incidents have led to heightened alerts and prompted a wave of action from the U.S. government. Congress has put forward legislation regulating AI in elections. The Federal Communications Commission requires disclosure and transparency of AI-generated content in political advertisements. State governments have adopted various measures to try to contain the misuse of AI in elections. The private sector also has come up with voluntary measures. Social media giants X, Facebook and YouTube have asked users to disclose AI contents on their platforms. However, because of the accelerating advancement of AI technologies, partisan gridlock on Capitol Hill and the commercial self-interest of the tech sector, the effort to rein in AI is facing multiple constraints.
Dialogue essential for governance
Given the shared interest in regulating AI, China and the United States have initiated an inter-governmental dialogue on the subject as an outcome of the San Francisco meeting between Joe Biden and President Xi Jinping in November.
With the growing concerns of AI disruptions in the political process, a China-U.S. dialogue is more relevant than ever. AI-enabled weapons of mass destruction would be devastating to humanity, but deepfakes and disinformation empower “weapons of mass distraction,” which mislead and manipulate the public, disrupt and destroy social cohesion, incite hatred and foment unrest.
While the U.S. is hyper-anxious over AI interference in its elections, China is concerned about malign actors using AI tools to stir up social unrest that undermines national security. Through an AI dialogue, the two countries could deepen their understanding of the other’s legitimate concerns, engage in mutual learning on best practices, cooperate to reduce risks and build confidence and make collective efforts toward designing frameworks of AI developments and applications.
In fact, the two countries have already aligned their efforts in the international arena, including signing the Bletchley Declaration at the first Global Summit on AI Safety in London in November, and supported each other’s AI resolutions in the UN General Assembly. These developments demonstrated that China-U.S. cooperation is essential to global AI governance.
Managing differences
However, the U.S. elections may also complicate any China-U.S. dialogue on AI.
First of all, given the heightened anxiety over cyber-interference, U.S. politicians have resorted to China-bashing. They accuse China of weaponizing AI to engage in “foreign malign influence” and to interfere in the elections.
Second, because the U.S. regards AI as a new frontier of strategic competition with China, it has been ramping up the repression of China in the “small yard” of AI, and erecting a “high fence” around it by tightening export controls on advanced semiconductors, expanding the entity list and restricting two-way investment in the Chinese high-tech industry.
OpenAI CEO Sam Altman’s recent op-ed in the Washington Post reveals that the sentiment of ideological confrontation has infiltrated the business sector. He frames China-U.S. AI competition as a battle between authoritarianism and democracy.
Last but not least, the two candidates for president seem to hold different views on AI governance. Vice President Kamala Harris, the Democratic Party’s nominee for the White House, is likely to continue the Biden administration’s regulation of AI, while she also has a long connection with Silicon Valley. She served as attorney general of California and then as a senator in Congress.
On the other side, former president Trump has vowed to rescind Biden’s executive order on AI “on day one,” and the Republican National Committee platform includes a plan to support AI Development “rooted in free speech and human flourishing.”. How the next U.S. administration will approach China-U.S. dialogue on AI hangs in the balance.
Given the high stakes, reining in AI is crucial for the benefit of both countries and the world. Talks must communicate mutual concerns, dispel misunderstandings and prevent this emerging technology from becoming a new source of tension. As major AI players, China and the U.S. can use dialogue to benefit themselves and contribute to safe and sustainable development of AI globally.
