Prompt Injection Risks in Compliance AI Systems
Compliance AI systems are uniquely vulnerable to prompt injection attacks.
Devraj Subramaniam
Staff Writer, AI & Emerging Compliance Technology
Devraj Subramaniam
Devraj Subramaniam holds a background in machine learning engineering and spent five years building risk-scoring models at two RegTech firms before shifting to research journalism in 2020. He tracks the deployment of AI agents and large language models inside compliance functions, scrutinizing vendor claims against real-world implementation evidence.
1 story