Deutsch English
← back to archive
AI & Technology · 538 words · MN-260708-0114

Anthropic, China, and the Tracking Code Excuse

MuckyNews reading

Anthropic has a bit of a control problem. To stop other companies from siphoning off their AI models — a practice known as "distillation" — the company embedded hidden tracking code into its coding tool, Claude Code. Fair enough to want to protect your intellectual property. The catch is that this tracking ran in the background, apparently without clearly informing users. Chinese authorities caught wind of this and issued a warning about a "backdoor" that could send sensitive data to remote servers without consent.

Anthropic’s response is essentially a classic dodge: users in China were never authorized to use Claude Code in the first place, they said. It’s a neat way to sidestep the actual issue — that a US tech company was secretly harvesting user data — by reframing the whole thing as a licensing dispute. It’s a tidy little micro-drama in the broader tech cold war: China yells "security risk," the US firm fires back with "you weren’t supposed to be on the platform anyway." But beneath the geopolitical noise, the core issue remains: hidden tracking code is a breach of trust, regardless of which side of the Pacific you’re sitting on.

Short version

  • What happened? Chinese authorities warned about hidden tracking code in

Anthropic’s Claude Code; Anthropic responded that Chinese users weren’t authorized to use it anyway.

  • Who is involved? Anthropic (US AI firm) and China’s Ministry of Industry

and Information Technology / National Vulnerability Database.

  • Why does it matter? It reveals how AI firms are secretly tracking users

to fight model theft, raising serious trust and privacy concerns.

  • What is still open? Whether the newer Claude Code versions are actually

clean, and how many users in China were genuinely affected.

  • What is the real pressure point? A company gets caught running hidden

surveillance code, and instead of offering transparency, falls back on a legalistic excuse about terms of service.

Perspective delta

hongkong_china_asia (SCMP): The SCMP coverage clearly frames this incident within the escalating US-China AI race. It emphasizes the "backdoor" risk and the potential transmission of sensitive data to US-based servers. Anthropic’s admission of embedding tracking code is presented as validation of Chinese security concerns. What gets downplayed is the fact that this was a countermeasure against product piracy; instead, the story is framed as a national security conflict. The institutional interest here is obvious: discourage Chinese users from relying on US software and paint American tech as an inherent security threat.

Source base / limits

The source base is extremely narrow. This article is based on a single report from the South China Morning Post. The perspective is therefore entirely filtered through a regional lens focused on the Chinese position. Anthropic’s statements are included, but only as presented within this one article. How US media or official US channels are framing the incident is completely absent from the material. The assessment remains one-sided and should be read as a very limited snapshot.

Sources / basis

  • MuckyNews container ID: MN-20260708-114
  • Source outlets/stations: South China Morning Post (SCMP)
  • Perspective groups: hongkong_china_asia
  • Article count: 1
  • Source count: 1
  • Perspective count: 1
  • Approximate raw text size: 1834 chars