Why is Trump pulling the plug on his AI Order? | DW News

Key Concepts

• Frontier Models: The most advanced, large-scale artificial intelligence models currently in development.
• Dual-Use Technology: Technology that has both civilian/economic applications and significant military/national security implications.
• The Brussels Effect: The phenomenon where EU regulations set global standards because multinational companies adopt them to maintain access to the European market.
• Cyber-Physical Risks: The potential for AI to facilitate sophisticated cyberattacks on critical infrastructure (e.g., power grids, hospitals, financial systems).
• Sovereign AI: The effort by nations to develop domestic AI capabilities to reduce dependence on foreign technology providers.

---

1. The Postponed Executive Order

President Donald Trump abruptly canceled a planned executive order aimed at regulating artificial intelligence. The primary justification provided by the President was a concern that such oversight would "dull the US's edge" in the global AI race against China. Trump emphasized that the US currently holds a lead in AI development and expressed fear that government intervention could act as a "blocker" to continued innovation.

2. Proposed Regulatory Framework

The draft executive order, as reported, aimed to establish a oversight mechanism for the AI industry:

• Pre-release Disclosure: Tech companies would be required to provide the government with an "early look" at new AI models before public release.
• Vulnerability Assessment: Government agencies would evaluate these models for dangerous capabilities, specifically focusing on cyber-security threats.
• Voluntary vs. Mandatory: While initially proposed as a voluntary notification process, experts suggest it could have served as a precursor to a formal government licensing regime for "Frontier models."

3. Case Study: Anthropic’s "Mythos"

The development of Anthropic’s model, Claude Mythos, served as a catalyst for the government's push toward regulation.

• Capabilities: Anthropic and independent evaluators (including the British AI Security Institute) identified that Mythos possesses the ability to perform sophisticated cyberattacks that were previously only possible by highly skilled human intelligence operatives.
• Risk Mitigation: Recognizing the danger, Anthropic opted not to release the model publicly, instead sharing it with a select group of private entities to "patch vulnerabilities" and strengthen defenses before a wider release.
• Strategic Shift: This event forced the White House to acknowledge that AI capabilities are advancing rapidly, necessitating a more assertive government role to prevent catastrophic accidents or exploitation by hostile actors (e.g., Russia, Iran, North Korea).

4. Geopolitics and the US-China AI Race

The summary of the competitive landscape is as follows:

• US Lead: The United States currently maintains a lead in AI, largely due to its dominance in the semiconductor supply chain (Nvidia, TSMC).
• China’s Strengths: While trailing in hardware, China possesses significant advantages in energy access, a large pool of talented researchers, and strong domestic tech firms.
• Military Integration: Both nations are racing to integrate AI into military targeting cycles and intelligence operations.
• International Regulation: There is currently very little binding international regulation, with the exception of narrow agreements between the US and China to prevent AI from controlling nuclear command and control systems.

5. The European Perspective

Europe continues to favor a more cautious, regulatory-heavy approach, exemplified by the EU AI Act. Former German Chancellor Angela Merkel argued that accountability for information dissemination is essential to protect the foundations of democracy. However, Europe faces a "difficult position":

• Resource Constraints: High energy costs and difficulty in building large-scale data centers have caused Europe to fall behind in the training of "Frontier models."
• Dependency: Middle powers and European nations are increasingly vulnerable to US policy shifts, as they may be forced to choose between US or Chinese technology, neither of which they fully control.

6. Notable Quotes

• President Donald Trump: "I think it gets in the way of... we're leading China. We're leading everybody. And I don't want to do anything that's going to get in the way of that lead."
• Angela Merkel: "Mistakes may be made, yes, but to believe that accountability for the dissemination of information is no longer necessary... will undermine the foundations of democracy."

---

Synthesis and Conclusion

The cancellation of the executive order highlights a fundamental tension within the US administration: the desire to maintain a competitive advantage in a high-stakes global race versus the urgent need to mitigate the national security risks posed by increasingly powerful AI models. While the US currently leads in hardware and model development, the "narrow" nature of this lead creates a policy dilemma. Heavy-handed regulation risks slowing domestic innovation, yet a lack of oversight leaves the nation vulnerable to cyber-physical threats. As the technology matures, the global community—particularly nations lacking domestic "Frontier" capabilities—will likely face increasing pressure to align with either US or Chinese regulatory and technological ecosystems.