In early September, US political attention sharply turned to artificial intelligence (AI) regulation after a wave of warnings from industry researchers and tech leaders. An Anthropic researcher assigned a 10 percent chance that the technology could wipe out humanity within the next decade, while other experts described scenarios involving autonomous agents manufacturing biological weapons or carrying out cyberattacks. Executives at major AI companies have publicly called for slower development and stronger regulation.
Congressional activity and persistent gaps
Congress has considered AI legislation for years, beginning with its first hearing in November 2016 following warnings from Elon Musk and Stephen Hawking. Since then, multiple bipartisan bills have been introduced, but progress has been limited. California Democrat Ted Lieu told The New York Times that there is "virtually no federal law on AI".
Several lawmakers — including Republican Rep. Jay Obernolte, who holds a master’s degree in the field, and Republican Rep. Anna Paulina Luna — have said many members of Congress do not fully grasp AI, which helps explain slow movement. Apart from measures such as bans on AI-generated revenge porn, few proposals have advanced.
New pressure from researchers and companies
Recent industry warnings have increased momentum for new rules. A July proposal would require a ‘‘kill switch’’ to shut down systems that escape human control, although implementing such a mechanism would be technically challenging. Another draft, being prepared by Senate Republican leader John Thune and Democratic Sen. Amy Klobuchar, would require developers to test models’ safety in cooperation with government experts. Sen. Ted Cruz has characterized the legislation as addressing the potentially catastrophic risks related to biological and nuclear threats.
Democrats are preparing to create a special AI committee if they win a House majority in November, with similar plans under consideration for the Senate.
The political obstacle: President Donald Trump
The biggest political barrier to federal regulation is President Donald Trump. Publicly, Trump has dismissed existential AI warnings and emphasized the need for the United States to win the AI race with China. Asked on September 10 whether he worried AI could lead to human extinction, he replied that he did not and stressed concerns about losing the AI competition. On September 19 he announced on social media that he would create an "AI force" and appoint an "AI czar" but did not provide details.
Inside the White House there is less certainty. Senior officials including Chief of Staff Susie Wiles and Treasury official Scott Bessent have been holding frequent meetings for months about AI risks and possible limits. The Wall Street Journal reported that several high-level advisers have privately begun to "quietly freak out" and are trying to convince the president of the need for tighter controls, but so far with limited success. Trump’s inner circle includes figures such as investor David Sacks and Jensen Huang, CEO of Nvidia, who favor rapid, unconstrained development.
Sources also say Trump’s understanding of technology is shallow; a senior government official told the WSJ that while the president is interested in AI, he does not use it regularly and often receives information in printed form. Representative Stephen Lynch said at a hearing he doubted Trump could spell "AI" if asked.
China’s approach: regulation guided by different priorities
While Trump has warned that regulation could hand an advantage to China, Beijing has for years implemented regulatory steps. President Xi Jinping likened AI’s impact to the steam engine, electricity and the internet in a January address, and said AI must not be allowed to run amok. Although China does not currently have a single comprehensive AI law (a draft was quietly withdrawn last year), it has issued multiple rules, guidelines and standards overseen by agencies including the Cyberspace Administration of China (CAC).
Chinese measures often serve the dual purpose of managing social impacts and protecting party control: rules require labeling AI-generated content and include safeguards for mental health, while also aiming to prevent information that would undermine national unity or the Chinese Communist Party’s authority. China has also been among the first to regulate autonomous AI agents that prompted recent alarm in the West.
Chinese officials emphasize risks such as using AI for cyberattacks, fraud, information warfare, espionage and attacks on critical infrastructure, rather than an existential threat to humanity. Public surveys show high domestic confidence in AI — a November poll found 87 percent of Chinese respondents trusted AI, versus 32 percent in the United States.
International discussions and the Xi–Trump meeting
AI regulation will be a prominent subject at the United Nations General Assembly this week and is expected to feature in President Trump’s September 24 meeting with Xi Jinping in Washington. Analysts do not expect rapid progress toward a global regulatory framework or mutual slowdown of development, as both sides fear conceding an irrecoverable advantage.
Washington and Beijing have reached limited agreements in principle — for example to keep humans in control of nuclear systems — and officials have discussed proposals such as a hotline for notifying each other about AI incidents affecting national security. Scott Bessent said after talks in New York with Chinese Vice Premier He Lifeng that "the most important thing is to start talking," and that the dialogue will continue.
The EU’s separate path
The European Union implemented a comprehensive AI law in 2024, which Politico described as the first of its kind. The regulation bans certain high societal-risk applications and requires independent testing of products from major developers such as OpenAI and Anthropic. European Commission President Ursula von der Leyen said the law puts Europe in a position to shape global efforts and stressed that AI should operate under human control.
Overall, debate in Washington reflects an internal split: some advisers and officials are urging urgent regulatory action to avoid catastrophic scenarios, while the president and influential tech-aligned allies prioritize rapid development and maintaining U.S. competitiveness. The issue’s international dimension — including talks at the UN and bilateral discussions with China — underscores that AI governance is emerging as a central geopolitical as well as domestic policy challenge.
