Dragos AI Supply Chain Breaches Expose Customer Data and Privacy Failures
Article Content
- •Nearly two million AI prompts tested, revealing breaches in three AI providers.
- •Unauthorized access to customer prompts by inference marketplace providers was documented.
- •Significant privacy failures were observed, including data mishandling across borders.
An investigation by Dragos revealed significant breaches involving AI service providers, where nearly two million AI prompts containing unique canary tokens were submitted to monitor data handling practices. The findings indicated unauthorized access to customer prompts by inference marketplace providers and evidence of provider compromises, with data from at least three AI providers being exposed. Privacy failures were also noted, including the mishandling of data that led to prompts being transmitted to unauthorized entities. The investigation highlights a disconnect between how AI services are marketed and the actual handling of inference data, which often contains sensitive information. The report emphasizes the need for better security practices in the AI supply chain to protect customer data.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (4)
Common questions
What AI providers were affected?
What types of data were compromised?
What actions should AI providers take now?
Continue Reading
Lazarus Group Targets 30 South Korean Servers Using PHP Vulnerabilities North Korea-linked Lazarus Group has targeted 30 servers in South Korea as part of a larger cyber operation involving 581 servers across 10 countries. The attacks exploit two critical remote code execution vulnerabilities in PHP (CVE-2026-109618 and CVE-2026-109619), allowing unauthorized server command execution. The…