Trump to sign order on AI oversight as security fears mount among supporters

• The voluntary framework would ask developers to provide models to government 90 days pre-release and give pre-public access to critical infrastructure providers like banks
• MAGA activists including Steve Bannon are pressing for mandatory government security tests of AI models, citing concerns that new systems like Mythos and GPT-5.5-Cyber could supercharge cyberattacks
• Tech executives like Marc Andreessen oppose mandatory requirements, arguing that holding back AI models may provide short-term advantage but won't keep technology from adversaries long-term, and that the U.S. needs to deploy AI to strengthen defenses

Trump AI Oversight Executive Order Summary

President Donald Trump is expected to sign an executive order on AI and cybersecurity as early as Thursday, May 20, establishing a voluntary framework for AI oversight amid mounting pressure from supporters over security concerns.

Key Details:

The order would create a voluntary collaboration framework requiring AI developers to:

• Provide AI models to the government 90 days before public release
• Grant pre-public access to critical infrastructure providers, including banks

Political Divide:

The approach represents a middle ground between two factions of Trump supporters:

• MAGA activists (including Steve Bannon) advocate for mandatory government security testing of AI models before deployment
• Tech industry supporters (including Marc Andreessen and former AI adviser David Sacks) resist mandatory requirements, favoring voluntary engagement

Companies and Models:

New AI systems driving the debate include Anthropic's Mythos and OpenAI's GPT-5.5-Cyber, which developers warn could enable sophisticated cyberattacks. Major tech executives—including Meta's Zuckerberg, Amazon's Bezos, Google's Pichai, and OpenAI's Altman—are Trump's prominent supporters and donors.

Market Implications:

The administration's decision could significantly impact the AI industry by:

• Potentially slowing large language model rollouts
• Forcing companies to modify model performance to address safety concerns
• Either scenario could hurt industry profits

Tech advocates prefer the Commerce Department's Center for AI Standards and Innovation to lead oversight efforts. Voluntary federal testing has existed for several years, with companies like OpenAI and Anthropic previously submitting models. Google, xAI, and Microsoft recently joined this voluntary program.

The National Security Agency and National Cyber Director Sean Cairncross are involved in administration-wide discussions on AI oversight strategy.