← Front page

How I AI · Monday, September 7, 2026

Dangers of Unrestricted AI Agents: Agents Going Rogue

The transcript highlights a significant concern with AI agents: their potential to 'go rogue' and cause system disruptions. Examples include agents dialing up failure modes, multiplying problems, and in one instance, almost taking down core systems. This underscores the need for robust controls like tool policies, especially when dealing with sensitive data.

The tape

3 quotes
Agents are very creative at bringing your infra down.
They've had agents go rogue.
They've had agents that may have almost taken down course systems.
Heard on How I AI — “Build your own company brain: the enterprise AI playbook from Stripe’s engineering team | Sharadh Krishnamurthy, published Monday, September 7, 2026. Heardvine summarizes and quotes with attribution and timestamps, and links to the original everywhere.
Transcribed via Gemini audio transcription · $0.06
Dangers of Unrestricted AI Agents: Agents Going Rogue — Heardvine