US government website used AI search tool from China that FBI said copied Anthropic

Read the original at The Straits Times ↗
The Straits Times · collected 2026-09-18 · by The Straits Times

Quick Summary

A US government website for the Federal Register used an AI search tool developed by Chinese tech giant Alibaba from September 16 until it was removed around the same day, according to Reuters. The FBI had previously accused Alibaba of “maliciously” copying technology from American firm Anthropic, raising tensions between the U.S. and China ahead of a high-level meeting. While US officials criticized Chinese AI companies for copying American technologies on an industrial scale, the use of Qwen by a federal entity sparked debate about the government's stance on employing Chinese-made AI tools.
Written locally by qwen2.5:14b on 2026-09-18, using this article's own text rather than the other coverage of the same event (that is the story summary below).

AI analysis runs on qwen2.5:14b, locally

Story summary

On September 16, a US government website, run by the National Archives, used Alibaba's AI search tool called Qwen for searching proposed federal regulations. This came despite an accusation from the FBI that Alibaba had "maliciously" copied technology from Anthropic, another AI company based in the United States. The Qwen-powered search was quickly removed after it became public knowledge. This incident occurred amid rising tensions between Washington and Beijing ahead of a planned meeting between their leaders. US officials have claimed that Chinese AI companies are systematically copying American technologies on an industrial scale, raising concerns about intellectual property theft and technological competition.

Written for “Chinese AI Tool Controversy” on 2026-09-18, grounded in this article and the 0 other(s) covering the same event.
Why this leaning score
The article's own words the score was based on. Each is quoted verbatim and was checked against the article text before being stored, so you can find it in the original.
Score +0.45 Confidence high
Leaning score +0.45 for article 17455 (high confidence, 2 verified quotes) · logged 2026-09-18

Signals How these are calculated →

Claims extracted
24
claim-shaped sentences
Uncertain
17%
4 of 24 hedged
Leaning
Leans right
of the writing, not the subject
Correction & hedging signals
58.9
corrections and hedging in what we collected; not a measure of accuracy
Outlets on this story
1
Technology
Narrative spread
1
articles carrying this framing
Analyzed 2026-09-18 · how these are computed

Story

📰 Chinese AI Tool Controversy
Technology · 1 article(s) covering the same event.

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Compared with similar articles

This article reads leans right and hedges 17% of its claims. Each row says how that neighbour differs.
BBC News
⚖️ leaning not scored 🔴 7% hedged 2 of 29 📰 publisher trust 96
“The articles discuss different topics related to AI and China-US relations, with Article A focusing on a critique of 'malicious competition' in AI development and Article B reporting on the US government's use of an AI search tool from a Chinese company that was accused of copying technology.”
NPR
⚖️ leaning not scored 🔴 23% hedged 6 of 26 📰 publisher trust 60
“The articles describe different events related to AI but occurring at distinct times and involving separate incidents.”
Semafor
⚖️ Leans right 🔴 0% hedged 0 of 4 📰 publisher trust 95
“The articles describe different events related to AI but are not reporting on the same specific incident.”
US AI regulation calls gather pace different event · 95%
Semafor
⚖️ Leans right 🔴 0% hedged 0 of 4 📰 publisher trust 95
“The articles discuss different aspects of AI regulation and use; one focuses on calls for legislation, while the other reports a specific incident involving a US government website using Chinese AI technology.”
CBS News
⚖️ Leans right 🔴 0% hedged 0 of 3 📰 publisher trust 71
“Article A discusses President Trump's comments about maintaining U.S. leadership in AI compared to China, while Article B reports on a specific incident where a US government website used an AI search tool from Alibaba that the FBI alleges copied Anthropic's technology.”
Daily Mail
⚖️ Leans right 🔴 14% hedged 4 of 29 📰 publisher trust 59
“The articles discuss different events related to AI and China, one about a warning by China's spy chief and another about a US government website using an AI tool from Alibaba.”
Semafor
⚖️ leaning not scored 🔴 3% hedged 1 of 32 📰 publisher trust 95
“The articles describe different events: one discusses AI companies' attitudes towards regulation, while the other reports a specific incident where a US government website used an AI search tool from China that was accused of copying another company's technology.”
Toronto Star
⚖️ Leans left further left than this 🔴 16% hedged 6 of 37 📰 publisher trust 61
“The articles discuss different aspects of AI development and use between China and the US, not the same specific incident.”
Semafor
⚖️ Leans left further left than this 🔴 0% hedged 0 of 4 📰 publisher trust 95
“The articles describe different incidents involving Anthropic and AI, one about declining a UK safety review and another about a US government website using an allegedly copied Chinese AI model.”
Semafor
⚖️ leaning not scored 🔴 10% hedged 1 of 10 📰 publisher trust 95
“Article A discusses a coalition urging the White House to release an AI framework, while Article B reports on a US government website using an AI search tool from China despite FBI concerns.”

Publisher

The Straits Times · 633 article(s) · 1 correction(s) detected
Running correction rate · 1 correction(s)
2026-09-13
Russia hits Ukrainian-Polish border area, Kyiv says

Who wrote this

The Straits Times
425 article(s) here · 1 carrying a prediction
🔮 Worsening the threat, open-source AI models can be run independently, allowing hackers to modify them or remove safeguards against cybercrime. “This gives malicious developers greater control over the model and more privacy, because they do not have to submit their source code to large cloud providers that may monitor their platforms for abuse,” Kamluk said.
🔮 The unusually direct rebuke from a Republican comes as the party controlling Congress and the White House faces growing concern that aggressive immigration enforcement could erode gains the president made among Latino voters in 2024.
🔮 Inside Microsoft and OpenAI, worry about damaging the publishing industry Karen Weise and Mike Isaac SAN FRANCISCO - Newly unsealed court documents showed considerable concern within Microsoft and its close partner OpenAI over the use of millions of news articles to develop artificial intelligence systems. As OpenAI was forging ahead with its work, Microsoft employees debated whether what OpenAI was doing represented the “largest theft of labour in human history” and could create a “doom loop” that could ultimately threaten the quality of the large language models they were building.
🔮 “On the other side of the Atlantic, we are seeing a change in political attitudes, a change in the assessment of the transatlantic alliance, which we might not have thought possible.”
🔮 Anthony James Kazmierczak pleaded guilty in May and admitted to attacking Omar because he disagreed with her progressive political views.
🔮 CSIS said two new communications compounds that had appeared at the base since mid-2024 likely form part of a broader intelligence and military command architecture, and would strengthen communications between Chinese forces operating in the Middle East and Africa and military headquarters in China. “The type and orientation of the base’s new equipment is varied, enabling transmission and detection across a wide range of signals and frequencies,” the report said.
🔮 Anthony James Kazmierczak pleaded guilty in May and admitted to attacking Omar because he disagreed with her progressive political views.
🔮 US said to delay excess capacity tariffs until after Xi summit AI generated The US is expected to delay the announcement of new tariffs over allegations of trading partners’ excess manufacturing capacity at least until after next week’s planned summit between Chinese President Xi Jinping and US President Donald Trump, according to people familiar with the matter.
🔮 Under the announced guidance, women were also set to be screened through a questionnaire and could then move to treatment for conditions related to hormonal imbalances.
🔮 The US government’s use of Qwen emerged amid a global reckoning about AI safety and a US debate over whether regulatory safeguards on AI are needed or would put the country at a competitive disadvantage.
More on this subject from The Straits Times
All 425 articles by The Straits Times →

Topics

Anthropic Chinese FBI Qwen WASHINGTON

Subjects

Chinese NORP · 9× Alibaba ORG · 4× FBI ORG · 4× Anthropic ORG · 3× WASHINGTON GPE · 3× China GPE · 2× Federal Register ORG · 2× Qwen ORG · 2× Reuters ORG · 2× the National Archives ORG · 2×

Narrative

The FBI last week accused Alibaba of copying Anthropic’s models to build its own artificial intelligence tools, raising tensions between Washington and Beijing ahead of a meeting next week between the countries’ leaders. US officials said then that Chinese AI companies were engaged in “malicious” copying of US technology on an “industrial scale”.
framing: assertive · carried by 1 article(s) · first seen 2026-09-18
🔮 The US government’s use of Qwen emerged amid a global reckoning about AI safety and a US debate over whether regulatory safeguards on AI are needed or would put the country at a competitive disadvantage.

Claims (24 extracted, 4 hedged)

US government website used AI search tool from China that FBI said copied Anthropic AI generated asserted
FBI → use → AI
A US government website employed an AI model from a Chinese company for users to search proposed federal regulations despite the FBI accusing the model’s developer, Alibaba, of “malicious” copying of US rival Anthropic’s technology. asserted
FBI → employ → technology
The website of the Federal Register, which is run by the National Archives, had the search tool using Alibaba’s Qwen model available on Sept 16 along with other search options. asserted
tool → run → options
It was unclear when it was deployed. uncertain
it → deploy → ?
The Qwen-powered search was taken down on Sept 16 around the same time social media posts appeared about it, according to screenshots and an archived version of the site’s source code reviewed by Reuters. uncertain
posts → power → Reuters
The FBI last week accused Alibaba of copying Anthropic’s models to build its own artificial intelligence tools, raising tensions between Washington and Beijing ahead of a meeting next week between the countries’ leaders. US officials said then that Chinese AI companies were engaged in “malicious” copying of US technology on an “industrial scale”. asserted
companies → accuse → scale
The FBI declined to comment on Sept 17. asserted
FBI → decline → Sept
The National Archives and the White House did not respond to requests for comment. asserted
Archives → respond → comment
A spokesperson for the Chinese embassy in Washington said the allegations of copying were “unfounded”. asserted
allegations → say → copying
The US government’s use of Qwen emerged amid a global reckoning about AI safety and a US debate over whether regulatory safeguards on AI are needed or would put the country at a competitive disadvantage. asserted
safeguards → emerge → disadvantage
AI experts said the use of the Chinese tool on the Federal Register’s website may not have posed national security risks. uncertain
use → say → risks
But it raised questions about whether the federal government was contradicting its own messages about Chinese AI models. asserted
government → raise → models
Daniel Castro, president of the Information Technology and Innovation Foundation, said there was a disconnect between the fierce US-China competition to build superior AI models and a US agency choosing to use a Chinese-made AI model. asserted
agency → say → model
“It’s one of the most insane things I’ve ever seen,” he said. asserted
he → ’ → things
The Federal Register website, which provides a daily list of newly proposed regulations, offered the Qwen model as a tool to browse public comments. asserted
which → provide → comments
Qwen is an open-weight model, meaning key components of the system are publicly available. asserted
components → mean → system
Developers can download and modify the model for specific uses, often at lower cost than closed models from companies such as Anthropic and OpenAI. asserted
Developers → download → Anthropic
“No federal government entity should use a Chinese AI model,” said US Representative John Moolenaar, a Michigan Republican who chairs the House China Committee. asserted
who → use → Committee
“Doing so only makes the federal government more dependent on Chinese AI models and that is not in the national interest”. asserted
that → do → interest
Security experts say open-weight models can run on a company’s own IT infrastructure, giving organizations greater control over sensitive data that might otherwise be shared with outside providers. uncertain
that → say → providers
Georgetown Law Professor Anupam Chander said the use of Qwen did not appear to present an immediate cybersecurity risk, though that depends on how the model was trained. asserted
model → say → risk
The Federal Register website’s content is already public, so the model was not working with sensitive government information. asserted
model → work → information
Democratic US Senator Mark Warner of Virginia, vice-chairman of the Senate Intelligence Committee, said in a statement that the risk depends on whether US data left the government’s security boundary and was processed on Alibaba-controlled systems. asserted
data → say → systems
In April, the US House committees on China and homeland security sent a letter to Airbnb, which also uses Qwen, asking for details about its use as part of an investigation into the “national security risks created by the integration” of Chinese AI tools into platforms used by Americans. asserted
which → send → Americans
💬 Give feedback
🕘 History 🎫 Support