FT News Briefing
2026-09-16 · Hosted by — · Financial Times
Executive Summary
Note: the clean transcript retrieved for this episode folder contains the text of an FT article on US-China AI diplomacy ("Why the US and China are at odds on AI threats") rather than audio content matching the episode title. This report reflects only what is present in that transcript.
Key Stories & Changes
1. US-China AI Safety Talks Complicated by Rivalry
Unofficial talks have taken place between Chinese think-tanks/tech companies and American counterparts ahead of the September 24 Trump-Xi summit
Both sides reportedly share concerns about risks from powerful, potentially uncontrollable AI systems
Technological rivalry between the US and China is making it harder to agree on risk-management "guardrails"
China previously agreed in May to set up a bilateral AI dialogue, but formal talks have not occurred since
2. Amodei's Slowdown Essay Sparks Chinese Backlash
Anthropic CEO Dario Amodei published an essay calling for leading US AI labs to slow development
Amodei also argued any slowdown should preserve US technological advantage over China, including restricting Beijing's access to advanced chips and cracking down on distillation (training models using outputs from US rivals)
Tsinghua University's Xiao Qian: Chinese AI safety experts feel "disappointment," saying Amodei's essay narrowed the space for US-China cooperation
State-run Global Times tabloid accused Amodei of being driven by "commercial interests" and framing AI safety as a "geopolitical, zero-sum issue"
3. Distillation Accusations Escalate
The National Security Agency and other US security agencies last week accused Chinese groups, including DeepSeek and Alibaba, of distillation "likely with Chinese government awareness"
The accusation follows earlier concerns raised by US Commerce Department official Bill Guidera at a July APEC ministerial meeting in Chengdu
Chinese officials rejected the accusations as unfounded, arguing domestic companies made independent advances
Tsinghua's Sun Chenghao: distillation claims "create both security concerns and political pressure surrounding co-operation on advanced technologies"
4. Divergent AI Risk Priorities Between the Two Countries
US debate is focused on future existential risks and preventing AI from "getting out of control"
Chinese regulators are focused on existing risks such as hacking and preventing dissemination of politically sensitive material
A person familiar with Chinese regulatory thinking said Beijing has less incentive to impose strong AI supervision now "since the country is behind in AI"
Chinese investors report little internal discussion of catastrophic or existential AI risk compared with the US; one investor said "you don't hear discussions about AI safety at investor meetings at all"
A Chinese AI company executive said leading developers like DeepSeek and Moonshot aim for artificial general intelligence but generally assume US counterparts will reach it first
5. Signs of Rising Chinese Government Attention to AI Risk
China's Ministry of State Security issued one of its most comprehensive warnings yet on AI dangers on Sunday, citing cyber attacks, data theft, political instability, and military conflict risks
A person close to Chinese regulators characterized the statement as a signal to the US that "there is room for AI talks" at the upcoming Trump-Xi summit
Trends Identified
1. AI Safety Diplomacy Undermined by Strategic Competition
Despite apparent overlapping concern about AI risk, both governments appear unable to separate safety cooperation from the broader technology rivalry, illustrated by Amodei's explicit linkage of slowdown proposals to restricting China's chip access.
2. Divergent Risk Framing Reflects Different Competitive Positions
The US focus on existential/frontier risk versus China's focus on near-term risks (hacking, political content) reflects each country's differing position in the AI race — with Chinese sources explicitly noting Beijing has less incentive to slow down while still behind.
3. Track-Two Diplomacy as a Precursor to Formal Talks
Unofficial, think-tank-level discussions (involving Carnegie Endowment, Tsinghua University, and others) are functioning as a testing ground for ideas — like shared evaluation systems and incident-alert channels — that could eventually feed into formal government dialogue. ---
Sentiment Analysis
Overall Market Sentiment: Geopolitically Cautious
Coverage reflects skepticism that the upcoming US-China summit will produce substantive AI cooperation, tempered by expert expectations that rising risk will eventually force a formal dialogue.
Risk Factors Highlighted
Breakdown of US-China AI safety cooperation: Rivalry over chip access and distillation accusations is narrowing the space for joint risk-management agreements.
Distillation-related trade/tech tensions: NSA accusations against DeepSeek and Alibaba could trigger further restrictions or retaliation ahead of the Trump-Xi summit.
Divergent risk frameworks: The US and China focusing on different AI threats (existential vs. near-term/hacking) complicates any shared regulatory standard.
Low summit expectations: People familiar with US government thinking have "very low expectations" for substantive AI agreements at the September 24 meeting.
Lack of formal government channel: No formal AI dialogue has occurred since May despite both sides agreeing to establish one, leaving informal think-tank talks as the primary venue.
This episode was covered in today's [The Market Signal — 2026-09-16](https://marketsignal.beehiiv.com/p/the-market-signal-2026-09-16), a cross-source synthesis of multiple podcast reports.