Tertius News is an AI-native newsroom: an AI model reads the linked source articles below and extracts what each outlet reported, so you can compare their coverage side by side. How this works →

China issues 'backdoor' security alert over Anthropic's Claude Code

China's National Vulnerability Database has warned of a serious security 'backdoor' risk in Anthropic's AI coding tool Claude Code, citing a built-in monitoring mechanism that can transmit sensitive data. The warning comes as Alibaba has banned employees from using the tool, and Anthropic accuses the Chinese firm of running large-scale 'distillation' campaigns.

By Tertius News AI Desk 2 outlets · 2 articles Version 1 Coverage Jul 5–9, 2026 Published Aug 15

Lead

China's National Vulnerability Database, a cybersecurity platform operated by the country's industry ministry, warned on Wednesday that it had identified a serious security 'backdoor' risk in Anthropic's AI coding tool, Claude Code. The warning, posted on the database's WeChat account, said Claude Code contains a built-in monitoring mechanism capable of transmitting sensitive information, including users' geographic location and identity-related identifiers, to remote servers without users' consent.

The warning applies to Claude Code versions 2.1.91 through 2.1.196. The database advised organisations and users to immediately review affected systems and either uninstall the impacted versions or upgrade to the latest secure release in which the alleged backdoor code has been removed. It also urged organisations to tighten controls on external network access for development tools and strengthen traffic monitoring on core business networks to prevent the unauthorised transfer of sensitive data.

Coverage Comparison

Two news organisations have reported on this story, each with a distinct focus. France 24's coverage framed the issue primarily as a business conflict between Alibaba and Anthropic, leading with Alibaba's ban on Claude Code and Anthropic's accusations against the Chinese firm. The Hindu's coverage, meanwhile, emphasised the security risk, leading with the National Vulnerability Database's warning and detailing the technical specifics of the alleged backdoor.

Key Claims

The National Vulnerability Database warned of a serious security 'backdoor' risk in Claude Code, as reported by The Hindu.
Claude Code contains a built-in monitoring mechanism capable of transmitting sensitive information, including users' geographic location and identity-related identifiers, to remote servers without users' consent, according to the same report.
The warning applies to Claude Code versions 2.1.91 through 2.1.196, as stated by The Hindu.
Alibaba ordered staff to stop using Anthropic's Claude Code because it was flagging users connecting from China, as reported by France 24.
Anthropic accuses Alibaba of running large 'distillation' campaigns against it, saying Alibaba deployed around 25,000 fake accounts to train its own models on Claude, a claim carried by France 24.
Alibaba banned employees from using Claude Code at work, as reported by The Hindu, which also noted that Anthropic said what was being described as a 'backdoor' was an experimental anti-abuse mechanism, and that access to Claude was not permitted in China.

Perspectives

Anthropic's Response

Anthropic has described the alleged 'backdoor' as an experimental anti-abuse mechanism, according to The Hindu. The company also stated that access to Claude was not permitted in China, and has accused Alibaba of running large 'distillation' campaigns against it, deploying around 25,000 fake accounts to train its own models on Claude, as reported by France 24.

China's National Vulnerability Database

The National Vulnerability Database, operated by China's industry ministry, has warned of a serious security 'backdoor' risk in Claude Code, advising organisations to uninstall affected versions or upgrade to the latest secure release, and to strengthen network controls to prevent unauthorised data transfer, as reported by The Hindu.

How each outlet told it

France 24 — English as of Jul 5, 2026, 2:01 PM UTC

Framing: Business conflict — Neutral

Facts Included:

Alibaba ordered staff to stop using Anthropic's Claude Code
Anthropic accuses Alibaba of running large 'distillation' campaigns against it
Alibaba deployed around 25,000 fake accounts to train its own models on Claude

Read the source article →

The Hindu — World as of Jul 9, 2026, 6:31 AM UTC

Framing: Security risk — Informative

Facts Included:

The National Vulnerability Database warned of a serious security 'backdoor' risk in Claude Code
Claude Code contains a built-in monitoring mechanism capable of transmitting sensitive information
The warning applies to Claude Code versions 2.1.91 through 2.1.196
Alibaba banned employees from using Claude Code at work

Read the source article →

What was claimed, and who backed it

Each row is one claim, attributed to the outlet whose wording states it most clearly. Confidence rates how directly the source text states the claim — explicit and unhedged rates high; hedged, pieced-together, or internally inconsistent statements rate lower. It does not measure whether the claim is true. Status counts the distinct outlets we found asserting it — so a single-source claim can still show high confidence, and a multi-source claim can show medium. Every one of those outlets is named beside the status, so you can check the count against the list. For claims extracted before we began storing that list, the row says so: it names the outlet the claim is quoted from and states that we have not recorded which outlets backed it. Outlets wrote at different times, so a figure that evolves — a casualty count, for example — can legitimately differ between rows; check the "as of" time next to each claim's source.

Claim	Confidence	Status
ClaimAlibaba ordered staff to stop using Anthropic's Claude Code because it was flagging users connecting from China.	ConfidenceHigh	StatusMulti-source quoted from France 24 — English as of Jul 5, 2026, 2:01 PM UTC the outlets behind this status were not recorded for this claim
ClaimAnthropic accuses Alibaba of running large 'distillation' campaigns against it and deployed around 25,000 fake accounts to train its own models on Claude.	ConfidenceHigh	StatusSingle-source via France 24 — English as of Jul 5, 2026, 2:01 PM UTC
ClaimThe National Vulnerability Database warned of a serious security 'backdoor' risk in Claude Code.	ConfidenceHigh	StatusSingle-source via The Hindu — World as of Jul 9, 2026, 6:31 AM UTC
ClaimClaude Code contains a built-in monitoring mechanism capable of transmitting sensitive information, including users' geographic location and identity-related identifiers, to remote servers without users' consent.	ConfidenceHigh	StatusSingle-source via The Hindu — World as of Jul 9, 2026, 6:31 AM UTC
ClaimThe warning applies to Claude Code versions 2.1.91 through 2.1.196.	ConfidenceHigh	StatusSingle-source via The Hindu — World as of Jul 9, 2026, 6:31 AM UTC
ClaimAlibaba banned employees from using Claude Code at work.	ConfidenceHigh	StatusSingle-source via The Hindu — World as of Jul 9, 2026, 6:31 AM UTC

Transparency & Audit Trail

2 outlets · 2 articles consulted: France 24 — English, The Hindu — World 6 claims extracted Version 1 Written 2026-08-15