0
Applied AI·November 21, 2026·1 min read

Nearly all software teams admit to shipping AI-generated code as study finds verification 'has not kept pace with that shift'

Share

Org charts have quietly accepted AI as a co-author while keeping 2015-era review processes—so unknown, unverified code is now in production everywhere. If you’re a CTO, your near-term risk isn’t AGI, it’s that your verification bandwidth and tooling lag your generation volume by an order of magnitude.

Applied AI

Researchers used canary credentials in requests to AI routers like OpenRouter and Vercel, even under "ZDR" policies, and saw "privacy" often broken, keys stolen

The AI routing layer is now a concrete data exfiltration risk, not an abstract privacy concern—2M planted tripwires lighting up means secrets are already leaking across the supply chain. If you're sending credentials, PII, or proprietary prompts through third-party routers, treat them as hostile networks and lock down what ever leaves your VPC.