US and Chinese officials are racing to establish formal channels for managing AI risks as President Donald Trump prepares to host Chinese President Xi Jinping in Washington. Following talks between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng in New York, the two sides agreed to create a US-China AI dialogue framework, including a notification mechanism for AI incidents tied to national security. Bessent described the discussions as very successful, with another meeting expected in about two months, possibly in Shenzhen.
AI has moved to the center of bilateral relations, overtaking Taiwan as the most pressing issue for both governments. The upcoming Xi-Trump summit on September 24 reflects a shared urgency over catastrophic miscalculation from AI in defense and nuclear systems. For Trump, a deal on AI offers a way to offset pressure from high energy prices. For China, the concern is that unchecked AI autonomy could threaten party control. Deep distrust still complicates any path toward agreement.
Meanwhile, China is expanding the physical infrastructure behind its AI ambitions. Dozens of new data centres are under construction in Inner Mongolia as Beijing seeks to lead the world in artificial intelligence. The build-out highlights how differently people in China and the US view the data centre industry, especially with the Xi-Trump meeting approaching.
On the domestic front, US President Donald Trump announced plans for an AI Czar position to coordinate policy amid intensifying competition with China. California Governor Gavin Newsom separately proposed a commission to regulate technology companies. Illinois Governor JB Pritzker and Newsom have both issued executive orders fast-tracking AI safety measures, acting within just four days according to the Transparency Coalition.
In other developments, a UN Women report launched on September 22 reveals that men hold nearly 80% of AI jobs and almost 90% of senior leadership roles in technology. Women face heightened risks from AI including technology-facilitated violence and gender discrimination. Sarah Hendriks of UN Women stressed that women must be equally involved in designing and regulating AI systems.
AI-generated videos depicting violence against women are spreading across social media, experts warn. One fake clip of a football reporter being struck went viral after a player commented on it. These range from humorous fakes to deeply realistic and harmful content that reinforces beliefs leading to real-world abuse.
On the regulatory side, the New York Department of Financial Services warned that AI governance rules must cover agentic AI. About 76% of companies are already using or planning to use agentic AI, and the Defense Department will soon audit defense contractors on compliance. Twenty state privacy regulators can also enforce these rules, meaning agents should be treated like a second type of identity with the same access rules and logging as human users.
Mathematicians have turned to AI to solve a decades-old problem in algebra. A team from Caltech, MIT, Temple University and other institutions used AI tools to find the polynomial for the M23 group, one of 26 sporadic groups that had resisted solution, within three months. Kyu-Hwan Lee noted that this kind of work was not really possible five years ago.
In the travel industry, Expedia vice president Daniel Shin Un Kan said AI agents are becoming gatekeepers for travel brands. He urged companies to ensure AI acts as an ambassador for customers rather than just for themselves, emphasizing that AI needs to benefit travelers.
A debate is also unfolding over how AI risks are discussed. AI CEOs are promoting doomsday scenarios about humanity's extinction by the end of the decade, and hosts on the Always at War podcast, including Evan Hubinger of Anthropic, question whether this fear-based focus distracts from real harms AI is causing right now.
Key Takeaways
- <ul><li>US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng agreed to a US-China AI dialogue framework with a notification mechanism for AI incidents and national security issues after talks in New York on September 20.</li><li>Chinese President Xi Jinping is expected to visit the White House on September 24, with AI now the top priority over Taiwan as both sides seek to reduce catastrophic miscalculation in defense and nuclear systems.</li><li>About 76% of companies are already using or planning to use agentic AI, and the New York Department of Financial Services requires AI governance rules to cover these systems, with the Defense Department soon auditing defense contractors.</li><li>Men hold nearly 80% of AI jobs and almost 90% of senior leadership roles in technology, according to a UN Women report launched on September 22 at the UN General Assembly.</li><li>Illinois Governor JB Pritzker and California Governor Gavin Newsom issued executive orders fast-tracking AI safety measures within four days, according to the Transparency Coalition.</li><li>Trump announced the creation of an AI Czar position, while California Governor Gavin Newsom proposed a commission to regulate tech companies, as the US and China compete for AI dominance.</li><li>China is building dozens of new AI data centres in Inner Mongolia ahead of the Xi-Trump meeting, seeking to lead the world in artificial intelligence.</li><li>Mathematicians from Caltech, MIT, Temple University and other institutions used AI to solve the inverse Galois problem for the M23 group within three months, a task not feasible five years ago.</li><li>AI-generated videos showing violence against women are spreading on social media, with experts warning they reinforce harmful beliefs that lead to real-world abuse.</li><li>Evan Hubinger of Anthropic joined the Always at War podcast in questioning whether AI CEOs' focus on doomsday narratives distracts from real harms AI is causing now.</li></ul>
US and China plan AI safety notification mechanism after New York talks
US Treasury Secretary Scott Bessent met Chinese Vice Premier He Lifeng in New York on September 20. The two sides discussed creating a US-China AI dialogue framework with a notification mechanism for AI incidents and national security issues. Bessent called the talks very successful and said the countries would probably meet again in about two months. China described the exchange as candid and constructive.
The narrow path to a Xi-Trump AI deal
Chinese President Xi Jinping is expected to visit the White House on September 24. AI is now seen as the priority over Taiwan because both sides want to reduce catastrophic miscalculation from AI in defense and nuclear systems. Trump faces pressure from high energy prices and may need a win. China is also concerned about AI autonomy threatening party control. Any agreement would need to overcome deep distrust between the two powers.
Beyond the trade truce, AI becomes key to China-US risk management
As Donald Trump and Xi Jinping prepare to meet in Washington, AI is emerging as a major issue. After talks between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng, the sides agreed to a formal AI dialogue and plan to meet in Shenzhen in about two months. They also discussed an incident line for national security-level AI events. AI is both a source of rivalry and a shared risk that could affect the whole international system.
China and the US must manage AI risks together
AI risks are no longer hypothetical for China and the United States. Chinese President Xi Jinping's upcoming US visit will focus on managing growing AI competition. The US has pushed an aggressive AI strategy and says it must win the AI race against China. China has surpassed the US in highly cited AI papers and leads in AI patents. Both countries are advancing military uses of AI, increasing the need for cooperation on shared risks.
Expedia exec says AI should be an ambassador in travel
Daniel Shin Un Kang, Expedia's vice president of global product development, spoke at the Travel Technology Europe conference in London. He said AI agents are becoming gatekeepers for travel brands. He urged companies to make sure AI works as an ambassador for customers, not just for the company. He stressed that AI needs to be used in ways that benefit travelers.
AI CEOs push doomsday narrative while real dangers go unnoticed
AI CEOs are pushing the idea that artificial intelligence could end humanity by the end of the decade. Hosts on the Always at War podcast question why leaders focus so much on this extreme scenario. The article suggests this fear-based talk may distract from real harms AI is causing right now. Evan Hubinger of Anthropic is among those involved in the discussion.
Women left behind as men dominate AI jobs and leadership
A UN Women report launched on 22 September shows men hold nearly 8 in 10 AI jobs and almost 9 in 10 senior leadership roles in technology. The report was presented at the United Nations General Assembly in New York. Women face greater risks from AI including technology-facilitated violence and gender discrimination. Sarah Hendriks of UN Women said women must be equally involved in designing and regulating AI systems.
Trump announces AI Czar as US weighs regulation and China rivalry
US President Donald Trump announced he will create an AI Czar position to address concerns about artificial intelligence. He said only high-IQ individuals need apply. The move comes as California Governor Gavin Newsom proposed a commission to regulate tech companies. The US and China are competing for AI dominance, and discussions about sharing information on AI technology are underway. US Finance Minister Scott Bessent stressed the importance of transparency between the two AI powers.
Mathematicians use AI to solve decades-old symmetry problem
Mathematicians used AI to solve the inverse Galois problem, a decades-old mystery about polynomial equations and symmetries. The team included researchers from Caltech, MIT, Temple University and other institutions. They focused on the M23 group, one of 26 sporadic groups that had resisted solution. Using AI tools, the team found the polynomial for M23 within three months. Kyu-Hwan Lee said this kind of work was not really possible five years ago.
Illinois and California governors fast-track AI safety measures
Illinois Governor JB Pritzker and California Governor Gavin Newsom have issued executive orders to fast-track AI safety measures. The actions address dangers of unchecked AI models. The Transparency Coalition reported that both governors acted within four days. The organization will monitor this developing story and provide updates as more information becomes available.
China builds AI data centres in Inner Mongolia
The BBC's Laura Bicker visits Inner Mongolia to see dozens of new data centres being built for China's AI goals. China wants to lead the world in artificial intelligence. The report looks at how people in China and the US feel differently about the data centre industry. This comes ahead of a meeting between leaders Xi Jinping and Donald Trump in Washington.
AI-generated videos of violence against women spread online
AI-made videos showing violence against women are spreading across social media. A fake clip of a football reporter being hit became widely shared after a player asked about it. These videos range from fake funny clips to very realistic and harmful content. Experts say this type of AI content helps support harmful beliefs that lead to real-world abuse against women.
Agentic AI must meet CMMC and state privacy rules
The New York Department of Financial Services warned on May 21, 2026 that AI governance rules must cover agentic AI. About 76 percent of companies are already using or planning to use agentic AI. The Defense Department will soon check defense contractors on this, and 20 state privacy regulators can do the same. The article says agents should be treated like a second type of identity with the same access rules and logging as human users.
Sources
- U.S. Treasury chief says Washington and Beijing plan AI safety notification mechanism; expert urges U.S. to attract Chinese investment for resilience
- The Narrow Path to a Xi-Trump AI Deal
- Beyond the Trade Truce: AI and the Next Phase of China–US Risk Management
- China, U.S. Must Manage AI Risks Together
- Expedia exec on making AI an ambassador in travel
- Is AI apocalypse talk distracting us from the real threat?
- AI is advancing fast. Women are being left behind in the workforce and leadership.
- 'AI force': Does Trump want to regulate AI amid growing concerns over artificial intelligence?
- Mathematicians use AI to find mysterious symmetries, solving decades-old problem
- Illinois, California governors fast-track AI safety measures with executive orders — Transparency Coalition. Legislation for Transparency in AI Now.
- Rare look at the remote data centres powering China's AI ambitions
- Holly Rowe isn't alone — AI-generated videos of violence against women are all over the internet
- Agentic AI Doesn't Get a Governance Exemption -Not Under CMMC, Not Under State Privacy Law
Comments
Please log in to post a comment.