Salt Security: AI Code Risks Worry 90% of Security Leaders

Nine out of ten security leaders are worried about the security risks from AI-generated code. This is according to new research from Salt Security. The study, which surveyed 100 IT security leaders, found that 67% of organizations are now widely using AI coding assistants. This shows how embedded AI has become in software development. However, governance frameworks are struggling to keep up. What's interesting is that 38% of organizations still rely on manual reviews for AI-generated code. Many security leaders feel this is unsustainable. The biggest risk identified by 29% of respondents was insecure coding patterns. Another 15% worried about generated code not aligning with internal security policies. AI coding assistants now generate almost half of all code on platforms like GitHub. Independent research also found that a lot of this AI-generated code has known vulnerabilities. Larger companies face more complexity as AI adoption grows. Organizations with over 500 employees are more likely to report challenges with governance consistency and developer over-reliance on AI. This matters because the rapid adoption of AI coding assistants is creating new security challenges for businesses.