Anthropic removes hidden tracker from Claude Code

Anthropic has removed a secret tracking feature from its Claude Code tool after a researcher exposed the code last week. The company had used the tracker since March to monitor users in China. Alibaba banned employee use of the tool in response.

A web developer known as Thereallo discovered the hidden code while researching privacy issues. The tracker used prompt steganography to send data on timezones, proxies, and potential links to Chinese AI labs without user detection.

Anthropic engineer Thariq Shihipar confirmed on X that the feature was an experiment to prevent account abuse and distillation attacks. He said the company had planned to remove it after developing stronger protections.

Last Friday Alibaba issued a memo banning Claude Code for work, citing security vulnerabilities from the tracker. The move follows accusations that Chinese firms have used distillation to copy US models like Claude.

Privacy advocates criticized the secret tracking. Thereallo stated that hiding the signal makes other privacy claims harder to believe.

Makala yanayohusiana

Illustration of US official revoking Anthropic AI access due to China concerns, showing models going offline.
Picha iliyoundwa na AI

Trump administration restricts Anthropic AI models over China ties

Imeripotiwa na AI Picha iliyoundwa na AI

The Trump administration directed Anthropic to revoke access to its advanced AI models following concerns about ties to China. The move led the company to take its most powerful offerings offline last week.

US artificial intelligence company Anthropic said users in China were never authorised to use its Claude Code product, responding after Beijing warned of security backdoor risks.

Imeripotiwa na AI

Anthropic announced updates to its Claude Design tool on June 17 that add design system imports, tighter integration with Claude Code, and new export options. The changes target enterprise teams needing AI-generated designs to align with brand guidelines and move smoothly into production.

The US government directed Anthropic to immediately suspend access to its Fable 5 and Mythos 5 AI models on Friday. The company complied with a full global shutdown after receiving the national security order at 5:21 p.m. ET.

Imeripotiwa na AI

Cybersecurity researchers have identified a fraudulent website mimicking the popular AI tool Claude that delivers backdoor malware to visitors. The discovery highlights how cybercriminals are capitalizing on growing interest in artificial intelligence platforms.

Jumanne, 23. Mwezi wa sita 2026, 15:15:48

Anthropic launches Claude Tag as Slack virtual coworker

Jumatano, 17. Mwezi wa sita 2026, 00:47:16

Anthropic pauses planned billing changes for Claude Agent SDK

Ijumaa, 8. Mwezi wa tano 2026, 18:37:26

Claude Mythos leak heightens cyber threats for banks

Jumatano, 6. Mwezi wa tano 2026, 03:25:03

Anthropic adds dreaming feature to Claude managed agents

Alhamisi, 23. Mwezi wa nne 2026, 14:42:22

Anthropic restricts Claude Mythos AI access via Project Glasswing amid hacking fears

Alhamisi, 23. Mwezi wa nne 2026, 11:03:42

Anthropic expands Claude AI connectors to lifestyle apps

Jumatatu, 20. Mwezi wa nne 2026, 20:41:39

Anthropic's Mythos AI model sparks hacking fears

Ijumaa, 17. Mwezi wa nne 2026, 19:24:12

Anthropic launches Claude Design research preview

Tovuti hii inatumia vidakuzi

Tunatumia vidakuzi kwa uchambuzi ili kuboresha tovuti yetu. Soma sera ya faragha yetu kwa maelezo zaidi.
Kataa