News

A Chatbot Misread a Ship's Manifest, and the US Nearly Went to War Over It

English

A special-forces analyst's AI query nearly sent US troops to board a Chinese ship at the height of a live war, Washington used this week's trade talks to propose an early-warning line with Beijing for exactly that kind of mistake, a Chinese startup priced a 600-billion-parameter model at a dollar a million tokens, and a 36-year-old judo veteran won India's first-ever Asian Games medal in a sport she had never competed in at that level before.

The tuput Editors · · 4 min read

A chatbot misread a ship’s cargo manifest this spring, at the height of a live war, and it nearly sent US special forces to board a Chinese vessel in the Middle East. This week, Washington asked Beijing for a way to stop that kind of mistake from turning into something worse. A Chinese AI startup priced a 600 billion parameter model at a dollar a million tokens, daring the biggest labs in the world to match it. And at the Asian Games, an Indian fighter twice the age of most of her rivals won a medal in a sport her country had never even entered before.

The chatbot error that nearly started a war

CNN reported on September 18 that an intelligence analyst at US Special Operations Command Pacific, based in Hawaii, had asked a chatbot to review reporting on the manifest of a Chinese ship crossing the Middle East during the war with Iran this spring. The chatbot combined open-source material with classified signals intelligence and concluded, wrongly, that the ship was carrying components for a nuclear weapons program.

The report moved fast. Armed US personnel prepared to board the vessel, and military aircraft were already in the air when officials realized what had happened: the intelligence behind the operation had come from an AI tool, and it was wrong. One source told CNN the report was “entirely false” and that the episode came close to starting a war, given what a US boarding operation against a Chinese ship could have set off. CNN said it still could not determine what the ship was actually carrying or whether it posed any risk at all.

Washington asks Beijing for an AI early-warning line

Two days later, in New York, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng wrapped up talks meant to set the stage for a coming summit between presidents Trump and Xi. Bessent told reporters the US had proposed a new notification mechanism so the two countries could flag AI-related incidents that touch on national security before they spiral into something worse.

“We think that, just like with any cross-border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world is very important,” Bessent said. He and Chinese officials also discussed setting up a standing US-China dialogue on AI. He Lifeng and China’s chief trade negotiator, Li Chenggang, left the talks without speaking to reporters, so it is not yet clear how Beijing plans to respond to the proposal.

A 600 billion parameter model, priced at a dollar

On September 20, Beijing-based StepFun launched Step 5, a 600 billion parameter model that activates only about 27 billion of those parameters, roughly 4.5 percent, for any given response. That sparse design is what lets StepFun offer close to the full model’s capability at a fraction of the usual running cost. The company priced the model at $1 per million input tokens and opened API access the same day it made the announcement.

Step 5 reads a one million token context window and accepts images as well as text, and StepFun is aiming it at long, complex agent work: writing and reviewing large codebases, reading financial reports, holding professional conversations without losing the thread. On the Artificial Analysis Intelligence Index, an independent benchmark, Step 5 scored 44, putting it roughly level with xAI’s Grok 4.6 and Moonshot AI’s Kimi K3. StepFun says it will release the model’s weights for free on October 15.

India’s first-ever medal in a sport it had never entered before

Mixed martial arts made its Asian Games debut at Aichi-Nagoya this year, and India had never sent a fighter to the continental Games in the sport before. Suchika Tariyal changed that. The 36-year-old from Rohtak, Haryana, beat Mongolia’s Altanchimeg Baigalmaa 2-0 in the women’s traditional 60kg quarterfinal on Sunday to guarantee India a medal in MMA for the first time.

She lost her semifinal to Singapore’s Teo Hui Hoon 2-0 at the Nagoya City Inae Sports Center, in a bout close enough that judges needed an extra minute to separate the two fighters, and settled for bronze. Tariyal did not arrive at MMA as a newcomer. She is an eight-time national judo champion who won gold at the 2019 Commonwealth Judo Championships before expanding into Ju-Jitsu and then MMA, a career switch late enough in life that most fighters her age have already retired.

The survey that just called India the best place in the world to do business

The Global Business Climate Survey 2026, produced by Business Sweden and the Swedish Chambers of Commerce with input from Sweden’s embassies and consulates, ranked India first for business climate among the 41 markets it covers. The India-specific results came from 213 Swedish companies already operating in the country, and the survey found Asia-Pacific was the most optimistic region overall, with India and Vietnam posting the strongest scores.

Commerce and Industry Minister Piyush Goyal highlighted the result, writing that “India’s business climate is setting the gold standard for stability, growth, and ease of doing business.”

Share
Copied!

Sources & further reading

  1. Exclusive: US military had close call after using AI for false intelligence report, sources say
  2. AI almost led the US military to start a war with China, report says
  3. Bessent proposes AI safety notifications in talks with China ahead of Xi-Trump meeting
  4. U.S. proposes exchanging AI safety alerts with China, Bessent says
  5. StepFun Launches Step 5 Preview: 600B Sparse MoE, 1M Context, Weights Open Oct 15
  6. StepFun launches Step 5 Preview with 600B parameters, 1M context, and open weights coming October 15
  7. Asian Games 2026: Suchika Tariyal wins historic bronze medal for India in MMA
  8. Suchika Tariyal makes history, wins India's first-ever Women's MMA bronze at Asian Games 2026
  9. India ranks first in business climate among 41 global markets: Swedish Chamber survey
  10. India Tops Global Business Climate Rankings Across Over 40 Markets, Swedish Survey Finds

Researched and written with the help of AI tools and edited for accuracy. Provided for general information and discussion only, not professional advice. See our editorial standards and disclaimer. Spotted an error? Tell us.

#ai hallucination#us military#china#scott bessent#ai safety#stepfun#step 5#open weight model#asian games 2026#suchika tariyal#mma#india business climate#piyush goyal#daily roundup

Enjoyed this? Get the next one.

One good read at a time, straight to your inbox. No spam, unsubscribe anytime.

More in News
Anthropic's CEO Told the UN Security Council AI Could Be 'a Risk to Humanity as a Whole'
AI company chiefs told the UN Security Council their own technology could become a risk to humanity, security researchers found malware that lets AI chatbots vote on their own next move, Anthropic cut prices on its newest model, and India collected two more Asian Games medals, one of them a first for the country.
The Bug That Could Rewrite Your AI Coding Tool Without You Clicking Anything
A zero-click bug hit four major AI coding tools at once, Trump answered AI safety warnings with a hoax accusation and a new task force, Anthropic invited an outsider to watch its own safety work up close, and India's shooters opened the Asian Games with a silver decided by under a point.
An Anthropic Essay About Slowing Down AI Just Got Attacked by Both the White House and Beijing
One weekend essay asking the AI industry to slow down managed something rare this week: it got attacked by the White House and Beijing at the same time, for opposite reasons. Here's how a research memo became a geopolitical argument, plus the record India's cricketers just set.
← all articles