# Lawmakers Tie Signers to AI Liability

A post argues that AI systems should be managed as non-human identities, with controls and restrictions comparable to those used for human accounts, and questions whether open-weight and closed-weight models are treated consistently as insider risks. It also asserts that lawmakers are placing liability on people who sign AI-related certifications, but identifies no specific law, agency action, contract, or implementation detail. For procurement teams, the post raises considerations for assessing AI access controls and certification exposure, but does not establish a new procurement requirement or opportunity.

- Procurement teams evaluating AI systems can assess how vendors control AI identities, restrict access, and address insider-risk concerns; the post presents this as a proposed governance approach, not a mandate.
- Organizations preparing AI-related certifications should identify the specific certification and applicable terms before signing; the signal does not name a certification or explain the claimed liability.
- The signal contains no solicitation, award, funding, or agency implementation details, so it does not identify a concrete contracting action.

**Jurisdictions:** federal
**Industries:** Information Technology
**Topics:** Artificial Intelligence, Cybersecurity
**Published:** October 16, 2026

### Key Quotes
> Truth be told, much like with any other non-human identity in your org, there should be controls and restrictions put around it. We've been doing that with human identities for decades now.
> — Original poster

### Sources
- [Nadella Calls Your AI an Insider Risk. Lawmakers Already Put Your Name on the Liability Certification.](https://www.reddit.com/r/cybersecurity/comments/1x7frxg/nadella_calls_your_ai_an_insider_risk_lawmakers) - reddit-cybersecurity