Report by Linux Foundation
State of AI Impact on Software Security and Cybersecurity 2026
Key Findings
Teams review transitive dependencies introduced by external packages about 50% of the time.
Only 57% of teams with an AI security policy require human review of AI-generated code.
58% of teams ship code faster than security reviews can keep up.
AI-assisted tools detect 30% of the vulnerabilities found in the past year.
60% of teams say AI coding assistants increase the number of external packages in their codebases.
69% of security practitioners say AI-assisted tools have meaningfully improved their security posture.
82% of security practitioners agree that overreliance on AI tools weakens developers' security skills.
AI tools make security work 17% faster on average.
False-positive vulnerability reports increase practitioners' workloads by 30%.
93% of security experts identify supply chain attacks as a leading concern.
Concern about AI-automated exploit generation and fuzzing exceeds practitioners' preparedness by 41 percentage points.
90% of security experts identify prompt injection as a leading concern.
30% of teams have a dedicated role responsible for AI-related security threats.
About one in three teams have a formal security policy that addresses AI.
75% of teams with a dedicated AI security role say they are well prepared for adversarial attacks on AI systems, compared with 50% of all teams.