The Trump administration is urged to conduct a security review before the release of AI models

robot
Abstract generation in progress

Investing.com — An advocacy group called on Monday for the Trump administration to conduct security threat assessments of advanced artificial intelligence models before they are publicly released, and to refuse government contracts for models that fail to pass the review.

The White House is currently responding to safety concerns raised by the Mythos model from Anthropic, which could make executing complex cyberattacks faster and easier, thereby creating national security risks.

The organization “Americans for Responsible Innovation” sent a letter to government officials, urging the Trump administration to establish mechanisms to review the cyberattack capability and weapon development potential of leading-edge models that major developers are set to roll out.

In the letter, the organization said that companies must pass the above review in order to qualify for government contracts.

The Center for AI Standards and Innovation (CAISI) is currently reviewing certain AI models through voluntary agreements with OpenAI, Anthropic, Google, Microsoft, and xAI.

The organization suggested that CAISI should take the lead in drafting mandatory requirements, and that Congress should set up a permanent enforcement agency within the U.S. Department of Commerce to enforce the related requirements.

MSFT-1.34% GOOGL+0.71%

Microsoft

Follow

Analyze MSFT

Included in our AI select strategy

·

View strategy details

415.12

▼-5.65(-1.34%)

Close · 09/05 · USD

412.75

▼-2.37(-0.57%)

Pre-market · 10:51:55

1D

1W

1M

6M

1Y

5Y

Max

Created with Highcharts 11.4.8 14:00 15:00 16:00 17:00 18:00 19:00 414 416 418

Contents

MSFT-1.34% GOOGL+0.71%

Analyze MSFT

The proposed requirements above would apply to companies that invest $100 million or more annually in frontier models trained on computing power, or companies with annual AI product and service revenues of $500 million or more.

California last year already set similar threshold standards for security reporting requirements.

This article was translated with the assistance of AI. For more information, please see our Terms of Use.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin