AI is ramping up coding velocity — and risk

This article analyzes the security implications of AI-assisted coding, arguing that higher developer throughput often produces fewer but larger multi-service pull requests that dilute reviews, expand blast radius, and contribute to a surge in security findings across dependencies, coding patterns, secrets, and cloud configurations. While experts dispute the magnitude of the reported increases, they converge on the need for strong context and guardrails (trusted images, policies), small and focused code reviews, proactive AppSec and intelligent agents in IDE/CI/CD, and clear governance to ensure accountability as AI accelerates development.