At the [un]prompted 2026 conference, TrendAI™ demonstrated how documents can be used to exploit AI-driven KYC pipelines and introduced FENRIR, an automated system for discovering vulnerabilities at scale. Principal Threat Researcher Sean Park highlighted the potential risks associated with document-based attacks on AI systems, while Threat Hunting Senior Manager Peter Girnus and Threat Researcher Demeng Chen presented FENRIR’s capabilities in uncovering weaknesses in AI ecosystems.
TrendAI™ at [un]prompted 2026: From KYC Exploits to Agentic Defense
Summary: TrendAI™ showcased the potential for AI-driven KYC pipelines to be exploited and introduced FENRIR, an automated system for discovering vulnerabilities in AI systems.
Key facts
- Documents can be embedded with hidden 'injects' to trick AI agents into reading and writing data across different customer records.
- FENRIR is an automated system that processes large codebases, identifies vulnerabilities, and generates exploit proofs for further review.
Why it matters
The findings from this research underscore the need for robust security measures to protect against potential attacks that exploit vulnerabilities within AI-driven processes. This is crucial as AI systems continue to play a significant role in various industries, including finance and cybersecurity.
@trendaisecurity
Embedded content for: TrendAI™ at [un]prompted 2026: From KYC Exploits to Agentic Defense