---
id: 1464
title: "Provisioning cloud infrastructure the wrong way, but faster"
source_domain: blog.trailofbits.com
source_url: "https://blog.trailofbits.com/2024/08/27/provisioning-cloud-infrastructure-the-wrong-way-but-faster/"
published_at: "2024-08-27 13:00"
tags: ["ai-security", "cloud", "cloud-security", "llm", "azure", "ai-generated-code", "trail-of-bits", "terraform", "credentials"]
phying_url: "https://phying.com/item/1464"
lang: en
---

# Provisioning cloud infrastructure the wrong way, but faster

## Summary

Trail of Bits shows blindly trusting LLM output for cloud provisioning—ChatGPT and Claude suggest hard-coded or pseudo-random passwords—yields working but insecure infrastructure.

## Intelligence Metadata

- **Sources:** blog.trailofbits.com
- **Published:** 2024-08-27 13:00
- **Original:** https://blog.trailofbits.com/2024/08/27/provisioning-cloud-infrastructure-the-wrong-way-but-faster/
- **Phying News:** https://phying.com/item/1464