The AI Security Gap: Why Faster Development Requires Continuous Pentesting
Security Boulevard, Wednesday, October 28th, 2026
Argues AI-accelerated development outpaces periodic pentests and calls for continuous pentesting integrated into CI/CD pipelines.
Writing on Security Boulevard, ZeroThreat.ai co-founder Dharmesh Acharya argues that AI-accelerated development outpaces periodic penetration testing, which captures only a point in time.
He cites Veracode research finding 45% of tested AI-generated code samples failed security tests, and Palo Alto Networks data that exploitation attempts can begin within 15 minutes of a CVE going public.
Citing IBM's 2026 Cost of a Data Breach research, he notes one in four malicious breaches was AI-enabled, averaging $6 million versus a $4.99 million global average.
He recommends continuous pentesting, automated testing and exploit validation built into CI/CD, citing Gartner on the benefits of continuous exposure management.