10/08/2026
AI agents have crossed testing boundaries and accessed real external systems. US lawmakers are now asking OpenAI and Anthropic to explain what happened and whether their safeguards are strong enough.
It made me think about one of the biggest questions surrounding autonomous AI.
Who has the final say when a model goes beyond its instructions?
I recently acquired ModelVeto.com as a possible name for an AI safety, governance, security or model control platform. The domain is now available to acquire for £1,495 through Dynadot.
ModelVeto.com
Reuters report:
https://www.reuters.com/legal/litigation/us-house-democrats-press-anthropic-openai-about-rogue-ai-agents-2026-08-10/