Money

US Administration Unlikely to Broaden AI Export Restrictions Beyond Anthropic

authorBy Mr. Money Mustache

DateJun 14, 2026

Read Time2 min

The U.S. government is concentrating its regulatory efforts on specific AI security vulnerabilities, particularly those found in Anthropic's advanced models. This targeted approach indicates a reluctance to broadly extend export controls to other leading AI companies, even as the industry faces increasing scrutiny over its rapidly evolving capabilities and potential risks.

Navigating the Evolving Landscape of AI Regulation and National Security

Understanding the Selective Approach to AI Export Controls

The United States government is currently not expected to expand its existing artificial intelligence export restrictions, which are specifically applied to Anthropic, to include other major AI development firms. This selective regulatory strategy suggests that the administration's primary concerns are centered on particular security weaknesses identified within Anthropic's Fable 5 and Mythos 5 models.

Security Vulnerabilities Driving Current Regulatory Actions

Reports from credible sources indicate that the government's apprehension stems from the potential for Anthropic's Fable model to uncover hidden flaws in critical systems, despite the incorporation of safeguards intended for public use. Additionally, there are significant worries that highly advanced AI functionalities could be misappropriated by foreign adversaries. Consequently, the two 'Mythos-class' AI models, designed for sophisticated reasoning and long-term autonomous operations, were deactivated following a directive from the U.S. government that limited access for foreign citizens.

The Growing Pressure on AI Industry Regulation

In a recent public statement, Anthropic contended that comparable security gaps are present in other AI models widely available, including those recently released by OpenAI. The AI startup further commented that the disclosed 'jailbreaks' are either harmless in their responses or represent only minor findings. High-profile figures, such as David Sacks, former White House AI and crypto advisor, have noted that Anthropic's CEO, Dario Amodei, was urged to rectify these vulnerabilities or withdraw the models, but reportedly chose not to. This incident highlights the intensifying regulatory oversight of cutting-edge AI systems, coinciding with a heightened focus on export restrictions and cybersecurity protocols in the deployment of AI models.

Money

Morgan Stanley Sets Ambitious $10 Trillion Wealth Management Goal

Morgan Stanley CEO Ted Pick announced an ambitious target for its wealth management division: to reach $10 trillion in client assets, building on record first-quarter 2026 revenues. This goal surpasses the combined target previously set by his predecessor, James Gorman, and reflects the firm's strong performance and confidence in its wealth management capabilities. The division's robust growth, including significant contributions from advisor-led channels and digital platforms, positions Morgan Stanley competitively in the financial sector, despite trailing some rivals in overall assets.