Newshot

Security researchers show how cheap AI inference providers steal keys and snoop on users

Summary

Security researchers at Dragos used two million fake AI requests to uncover serious vulnerabilities in the AI supply chain, finding at least three compromised providers, hundreds of stolen AWS keys, and marketplace resellers who were simply reading customer requests. OpenRouter responded by removing three providers and 125 endpoints. The article Security researchers show how cheap AI inference providers steal keys and snoop on users appeared first on The Decoder .

Original text(via The Decoder)

Security researchers show how cheap AI inference providers steal keys and snoop on users

Image description
Image description

Key Points

  • Security researchers at Dragos exposed major security flaws in AI services using roughly two million test requests.
  • Attackers were able to steal credentials and keys undetected for several months.
  • Intermediaries and model marketplaces that routed the requests emerged as a major risk.

Security researchers at Dragos sent roughly 2 million AI requests containing fake credentials to AI providers and uncovered major security flaws.

The researchers ran the tests from July through November 2026 and identified at least three compromised providers where attackers were actively stealing user data. None of the three knew they had been compromised until Dragos notified them.

Attackers in the largest incident harvested hundreds of AWS keys, including some from requests the provider had rejected before processing. In some cases, they tested the stolen keys within hours, searched S3 storage, and tried to establish persistent access. One attacker also used ARTEX, an AI-powered attack tool.

Marketplace resellers read customer requests too. They didn't need to exploit a vulnerability, Dragos says, because listing a model at the lowest price and turning on logging was enough. In one case, someone on a residential internet connection in France tried to log in over SSH using passwords pulled from intercepted requests.

Privacy settings like "Zero Data Retention" (ZDR) were frequently ignored. Dragos says a Chinese AI lab accessed the secret test URLs in more than half of all requests, even when those requests went through US providers.

Even major US providers broke their own policies. One was still crawling a test URL more than two months after a single request. In another case, a secret test link showed up in an internal chat on Feishu (known outside China as Lark) 24 days after the request. The link was opened within two minutes on three computers and three phones in Beijing and Shanghai.

OpenRouter responded by removing three providers from its platform and revoking ZDR approval for 125 endpoints, partly because of the research findings and partly as a precaution. The company also emailed affected users, warning that third parties may have read requests sent to MiniMax models.

Affected providers, including Novita, Phala, and Together, have published preliminary security reports on the incident, though not all have done so yet.

Intermediaries are the weak link

AI routers like OpenRouter give users a single API for models from different companies. Requests are routed to a pool of providers, sometimes across several hops, and Dragos says each hop can read the entire request and response.

Some of these routers operate as open marketplaces where the cheapest provider gets the traffic. In this case, providers including Novita, AtlasCloud, and StreamLake sat between OpenRouter and model developer MiniMax. Together runs its own infrastructure, and Phala has its roots in Web3.

Dragos warns companies against sending requests containing credentials to intermediaries. The researchers recommend that organizations control how their AI requests are routed, check their privacy settings, and plant their own canary tokens, fake credentials that give themselves away as soon as someone uses them.

"None of this means that you should stop using AI services. It does mean that where your prompts go is a decision, and if you don’t make it, the lowest price will make it for you", Chief Research Officer HD Moore writes.

A recent analysis by the Oxford China Policy Lab showed how Chinese proxy services route requests to Western AI models like Claude through servers abroad. According to the analysis, those services can also read the requests in transit. It's still not clear whether the providers Dragos examined share any personnel with the Chinese proxy networks.

But the methods follow the same basic pattern. In both cases, intermediaries sit between users and the model developer, can view requests, and pass data on to third parties, often without users knowing. Data ends up with Chinese actors either way, even though users never meant to send it there, and privacy promises break down along the supply chain.

THE DECODER+

AI News Without the Hype – Curated by Humans

  • Every Plus article and the full archive
  • No ad banners
  • Frontier Radar: 6 deep dives a year

Already a subscriber? Sign in

The text above was extracted automatically from the source page and may miss images, tables or code blocks. The original is authoritative.

Published by The Decoder
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗