中文翻译

摘要: Technology

White House Plans AI Safety Testing Framework for OpenAI, Anthropic Before Model Releases

A forthcoming executive order could create a voluntary government-AI industry framework for pre-d...

正文

Technology

White House Plans AI Safety Testing Framework for OpenAI, Anthropic Before Model Releases

A forthcoming executive order could create a voluntary government-AI industry framework for pre-deployment safety testing, with the NSA expected to play a key oversight role.

Bhaswati Guha Majumder

May 21, 2026 14:53 +08

White House plans a voluntary AI safety framework that would enable government agencies to pre-release test advanced AI models before deployment.

magnific.com

The White House is preparing a new executive order on artificial intelligence that could establish a voluntary information-sharing framework between the US government and leading AI developers to facilitate pre-deployment safety testing of advanced AI models.

According to reports, the order, which could be unveiled as early as this week, is expected to focus heavily on cybersecurity and national security risks posed by increasingly powerful

AI systems.

Under the proposed framework, AI giants such as

Anthropic

may voluntarily provide advanced models to government agencies for testing before public release, potentially up to 90 days in advance.

According to

, the draft order introduces a new classification called "Covered Frontier Models," which would identify advanced AI systems considered significant enough to warrant additional scrutiny before deployment.

The framework would reportedly involve multiple federal agencies in reviewing these models, while the National Security Agency (NSA) is expected to play a central role in classified evaluations.

"Any policy announcement will come directly from the president. Discussion about potential executive orders is speculation," a White House official

Nextgov/FCW, responding to reports surrounding the proposal.

Balancing Innovation and Security

The deliberations reflect growing tensions within the Trump administration over how aggressively

AI development

should be regulated.

While some officials have a

来源: Brave/ibtimes.sg

采集时间: 2026-05-21 20:16:24