Experts Urge Closing Identity Gaps as AI Reshapes Software Development Security
Summary
Security experts say AI-generated code and non-human identities are expanding SDLC risks. They recommend better identity visibility, short-lived credentials, activity logging, and closer security-development collaboration; AI may also speed vulnerability remediation.
Key points
- Experts say AI-generated code and a growing attack surface make it harder to embed security throughout the software development lifecycle.
- A cited Entro report found an average of 144 non-human identities for every human identity in enterprise environments.
- A cited Cloud Security Alliance study found 68% of organizations cannot distinguish actions performed by AI agents from those performed by humans.
- Experts recommend short-lived credentials and logging agent activity across its lifecycle.
- Roundtable participants urged security teams to collaborate with developers earlier and improve visibility into development tools with administrative access to pipelines or repositories.
- Participants said AI can help identify vulnerabilities and accelerate patching, even as attackers also use AI to find vulnerabilities or develop exploits.
Article Details
- Topic
- Identity security gaps in the software development life cycle, including risks from non-human identities and AI-generated code
People
Aftab BanthFormer global enterprise head of security for TikTok; discussed risks in source code and the software development life cycle, and AI-assisted remediation.Bobby BauerCISO of BlueFlag Security; discussed non-human identity risks and security-development collaboration.Renee GuttmannFounder of Cisohive; discussed integrating security into the software development life cycle and managing identity risks.Rusty PerryFormer VP of cybersecurity strategy at Western Union; discussed identity visibility and the use of AI in software development.