Trump could sign AI executive order as soon as Thursday
Trump’s Potential AI Executive Order Sparks Debate on Regulation and Cybersecurity
Trump could sign AI executive order – The Trump administration is preparing to take action on artificial intelligence (AI) regulation, with the possibility of issuing an executive order by Thursday. This move, according to insiders familiar with the developments, aims to establish a voluntary framework for government oversight of new AI models before they are released to the public. The order would require participating companies to share advanced versions of their AI systems with federal authorities for a specified period, as revealed by two sources close to the discussions.
While the details are still being finalized, the proposed executive order emphasizes collaboration between the government and private sector. A key aspect of the plan is the inclusion of a voluntary agreement that outlines the conditions for companies to submit their AI models for review. This agreement, as per the sources, would mandate a temporary handover of advanced models to the government, allowing for scrutiny prior to public launch. The initiative reflects the administration’s growing concern about AI’s potential to amplify cyber threats, a topic that has gained urgency in recent months.
Timeframe Disputes and Industry Preferences
One of the most contentious parts of the order is the proposed timeline for pre-launch evaluations. Industry leaders and federal officials have been engaged in heated debates over how long companies should be required to keep their models under government review. A draft version of the executive order suggests a 90-day window for assessments, but some AI firms, including prominent players like OpenAI and Anthropic, advocate for a shorter period, such as 14 days. These differing views highlight the tension between regulatory thoroughness and the need for rapid innovation in the AI space.
Anthropic, a key player in the AI sector, has been particularly vocal about the importance of timely oversight. The company unveiled its Mythos model earlier this month, which it claims can exploit cybersecurity vulnerabilities at an unprecedented speed. While the model remains unreleased to the public, Anthropic is offering early access to a select group of partners through its Project Glasswing initiative. This controlled distribution aims to allow businesses and government agencies to test the model’s capabilities and fortify their defenses against potential attacks.
Framework for Early Government Review
The executive order is structured into two main sections, each addressing distinct aspects of AI governance. The first focuses on cybersecurity, while the second, titled “covered frontier models,” defines the criteria for AI systems that would qualify for the voluntary review process. The latter section specifies which types of models—such as those capable of autonomous decision-making or real-time data processing—are eligible for early evaluation. This classification is designed to ensure that the most impactful AI technologies receive focused attention from regulators.
Under the cybersecurity framework, a clearinghouse will be established by the Treasury Department and other federal agencies. This clearinghouse will serve as a collaborative platform for AI developers and government officials to identify security risks and implement fixes in unreleased models. The proposal also calls for expanding the US Tech Force, a dedicated team of engineers tasked with modernizing federal computer systems. By increasing its workforce, the administration hopes to enhance the government’s ability to respond to emerging cyber threats.
Industry Collaboration and Strategic Access
OpenAI and Anthropic are among the major AI companies actively engaging with the White House on this initiative. Both firms have been working closely with federal representatives to shape the details of the executive order, ensuring it aligns with their interests while addressing national security concerns. This collaboration extends beyond Washington, with local and state officials also involved in the discussions. The goal is to create a cohesive strategy that leverages private sector expertise while safeguarding public infrastructure.
The executive order’s voluntary nature is intended to balance regulatory pressure with industry flexibility. Companies are not required to submit their models unless they choose to participate in the framework. However, this approach has sparked debate about its effectiveness. Critics argue that voluntary compliance may not be sufficient to address the scale of AI-related risks, especially as the technology continues to evolve rapidly. Supporters, on the other hand, emphasize that the order provides a starting point for dialogue, which is crucial for developing a long-term regulatory strategy.
Experts have noted that early evaluation can enable authorities to identify and mitigate potential threats before they manifest in real-world scenarios. For instance, advanced AI models are capable of accelerating cyberattacks by detecting patterns and exploiting weaknesses in systems that human analysts might miss. By engaging with the government prior to release, companies can help prevent these risks from escalating into full-scale breaches.
Historical Context and Recent Shifts
The Trump administration has historically favored a hands-off approach to AI regulation. This strategy allowed for rapid development and deployment of AI technologies without significant bureaucratic hurdles. However, recent events have prompted a change in direction. The introduction of the Mythos model by Anthropic, which reportedly outperforms existing AI systems in cybersecurity exploits, has raised alarms within the federal government. This model, still in a restricted access phase, has become a focal point for discussions on how to regulate AI’s impact on national security.
The Department of Commerce’s National Institute of Standards and Technology (NIST) had previously announced that major tech companies would share unreleased AI models for evaluation. This initiative, aimed at assessing the technology’s implications for public safety, was a significant step toward government involvement in AI oversight. However, the announcement is no longer visible on the Commerce Department’s website, suggesting that it may have been revised or replaced by the current executive order. The NIST’s role in this process remains a key area of interest for both industry stakeholders and watchdog organizations.
As the White House prepares to finalize the order, the focus is on creating a mechanism that allows for proactive threat detection. The voluntary framework is seen as a middle ground between strict regulation and complete deregulation, giving companies the autonomy to choose their level of participation. This balance is critical for maintaining the competitiveness of the US AI industry while addressing the growing concerns about security and transparency.
Industry insiders have expressed cautious optimism about the order. While some view it as a necessary step to align with global AI governance trends, others worry that the 90-day review period may be too lenient. The debate underscores the challenges of regulating a technology that is both transformative and rapidly advancing. Ultimately, the success of the executive order will depend on how effectively it fosters collaboration between the public and private sectors while ensuring that national security remains a priority.
The White House has not yet provided a formal statement on the matter, but insiders indicate that the order is close to finalization. This decision follows a series of high-profile AI developments that have prompted increased scrutiny from regulators. As the deadline for Thursday approaches, the administration is expected to present a comprehensive proposal that reflects both its regulatory philosophy and the evolving landscape of artificial intelligence.
