Widespread operational impact
“These technical vulnerabilities, if not tested, do not exist in isolation,” said Mindgard’s Garraghan. “They manifest as a wider organizational risk across the engineering field. The consequences of insufficient mapping of AI security tests directly lead to failures in security, security and business assurance when viewed from the perspective of operational impact.”
Sam Peters, chief product officer of compliance expert ISMS.Online, sees the widespread operational impact of Organziations’ tendency to ignore the trend of proper AI security review.
“When AI systems are flushed into production, we see repeated vulnerabilities in three key areas: model integrity (including poisoning and evasion of attacks), data privacy (such as training data leaks or unfortunately sensitive data) and governance gaps (from lack of transparency to poor access control),” he said.
