LLMs & Malicious Code Injections: 'We Have to Assume It's Coming'

A CISO roundtable at RSAC highlighted prompt injection into LLM-powered workflows as a rising organizational risk, citing scenarios where malicious prompting can generate socially engineered alerts and lead to unauthorized data sharing. Panelists underscored pervasive shadow AI, noted AI’s disruption of SIEM by reducing L1/L2 triage, and advised leaders to establish usage boundaries, deliver basic prompt-engineering training, and reinforce testing and continuous learning in application development.