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 →
Anthropic hits back after China warns of Claude Code ‘backdoor’ risks
China's National Vulnerability Database warned users of a security backdoor in Anthropic's Claude Code, urging uninstall or upgrade. Anthropic replied that Chinese users were not supposed to use the product, and confirmed it had embedded location-tracking code. Alibaba has reportedly banned Claude Code use from July 10.
China's cybersecurity platform, the National Vulnerability Database (NVDB), issued a warning on Wednesday about a security "backdoor" in Anthropic's AI coding tool, Claude Code. The database, affiliated with the Ministry of Industry and Information Technology (MIIT), advised affected users to immediately uninstall or upgrade to a secure version. Anthropic responded by stating that users in China were not supposed to be using the product in the first place, in a statement that came amid escalating US-China AI tensions.
Coverage Comparison
Dawn and the South China Morning Post both reported on the warning and Anthropic's response. Dawn emphasized the security warning, framing it as a "security backdoor" that could transmit sensitive information without consent. The South China Morning Post highlighted the US-China AI race angle, noting Anthropic's statement that Chinese users were not intended to use Claude Code and providing details on affected versions.
Both outlets reported the NVDB's recommendation for comprehensive checks and uninstall or upgrade, and noted that Anthropic blocks users and companies in China from accessing its products, though use may still occur via VPN or proxy services. The South China Morning Post additionally reported that Anthropic confirmed embedding hidden code to track user locations, a detail not mentioned in Dawn's coverage.
Key Claims
Security Warning: China's NVDB warned that Claude Code contains a security backdoor, posing a "severe threat" and potentially transmitting sensitive information, including user locations and identity-related identifiers, to remote servers without consent.
Affected Versions: The warning applies to Claude Code versions 2.1.91 to 2.1.196, which were released from April to late June. Newer versions have since been released.
Recommendations: The NVDB advised affected developers to immediately uninstall or upgrade to the latest secure version that has removed the backdoor code, and to strengthen control over external access permissions to prevent unauthorized transmission of sensitive data.
Anthropic's Response: Anthropic stated that users in China were not supposed to be using Claude Code, as the company blocks access from China and other adversarial nations. The company also confirmed last week that it had embedded hidden code to track user locations to stop illicit "distillation" of its models.
Alibaba's Ban: Chinese tech giant Alibaba reportedly told employees last week that the use of Claude Code would be banned from July 10 due to security concerns, according to people familiar with the matter. Anthropic has previously accused Alibaba of reverse-engineering its AI models in a process known as "distillation."
How each outlet told it
Dawn
Framing: Security warning — Neutral
Facts Included:
Chinese industry regulator warned users of a security backdoor in Anthropic's Claude Code
The alleged backdoor could allow the software to transmit sensitive information without users' consent
Claude Code is an AI coding agent that can generate computer code, debug software and review code based on user prompts
Anthropic blocks users and companies in China from accessing its products
The NVDB advised users to conduct a comprehensive check and uninstall or upgrade to the latest secure version
Alibaba banned the use of Claude Code due to security concerns
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
ClaimA Chinese industry regulator warned users of a security backdoor in Anthropic's Claude Code. The alleged backdoor could allow the software to transmit sensitive information without users' consent.
ConfidenceHigh
StatusMulti-source quoted from Dawnthe outlets behind this status were not recorded for this claim
ClaimClaude Code is an AI coding agent that can generate computer code, debug software and review code based on user prompts.