## White House Prepares EO Mandating US Agency Partnerships with AI Companies on Cybersecurity — Exempts Pre-Release Model Testing
The Trump administration is preparing to order US government agencies to formalize partnerships with artificial intelligence companies for cybersecurity purposes, according to sources familiar with the matter. The planned executive order would direct federal agencies to collaborate with AI firms on protecting government networks—but notably would not require government testing of AI models before their public release, a significant omission that signals the administration's preference for industry self-regulation over mandatory safety protocols.

The draft order, which has not yet been publicly announced, represents a consequential policy decision on how Washington will approach AI oversight in the national security domain. By mandating agency-AI company cooperation while declining to impose pre-release testing requirements, the administration is staking out a position that prioritizes speed of deployment and private-sector autonomy. The approach diverges from recommendations by some AI safety researchers and lawmakers who have called for more rigorous government evaluation of powerful AI systems before they reach users, particularly those deployed in sensitive government environments.

The move comes as competition intensifies over AI development and cybersecurity resilience, with the US seeking to maintain technological advantage while addressing vulnerabilities in critical infrastructure. The decision to forgo mandatory pre-release testing could accelerate AI integration into federal systems but raises questions about potential risks from inadequately vetted models operating in high-stakes government networks. The order's framework suggests the administration views AI companies as essential partners in defending against cyber threats rather than entities requiring heightened scrutiny—a posture that could reshape public-private dynamics in the sector while leaving open questions about accountability if deployed AI systems fail or are exploited.
---
- **Source**: Techmeme Echo RSS
- **Sector**: The Network
- **Tags**: AI, cybersecurity, executive order, regulation, US agencies
- **Credibility**: unverified
- **Published**: 2026-05-09 01:24:41
- **ID**: 80892
- **URL**: https://whisperx.ai/en/intel/80892