The U.S. government suspends access to Anthropic's advanced AI models

On Friday, June 12, at 17:21 ET, the artificial intelligence company Anthropic received a letter from the U.S. government ordering it to suspend access to its most advanced models, Fable 5 and Mythos 5, for any foreigner, including the company's own employees. The order gave no prior notice and no deadline for restoration.

Very few institutions had developed an operational dependence on these specific models: Fable 5 had been made public for less than a week, and Mythos 5 was accessible only to a small group of partners under Anthropic's Glasswing project. However, on July 1, the U.S. government lifted the suspension, allowing Anthropic to make these models globally available again.

But for institutions around the world, the short-lived order revealed how a foreign government can disrupt access to frontier AI capabilities in minutes, without notice and without recourse.

National security concerns

The government's official reason for the suspension was linked to national security concerns. According to Anthropic, the government discovered a "jailbreak" technique capable of bypassing Fable 5's security controls. Anthropic disputed the severity of the finding, calling it a "misunderstanding" and warning that applying the same standard across the sector would block the release of frontier models.

This dispute did not occur in isolation. A few weeks earlier, Anthropic had rejected the terms of the Pentagon's contract that would have allowed the military to use its models for any legal purpose, including autonomous weapons systems and domestic surveillance. After negotiations failed, the U.S. Department of Defense designated Anthropic as a "supply chain national security risk," a label historically reserved for foreign enemies, applied for the first time to a U.S. company. Anthropic is challenging this designation in federal court. Therefore, the company and Washington were already in an adversarial relationship before the export control directive arrived.

Impact on African institutions

African governments were not part of this dispute. They read the news Monday morning, three days after the order, along with everyone else.

However, this is not a story about a black Friday. Africa's dependence on foreign-controlled technologies is visible elsewhere: in telecommunications backbones, cloud infrastructure, and digital identity systems built and managed by U.S., European, and Chinese providers. Frontier AI now fits into the same pattern. The June suspension was not aimed at Africa or foreign institutions in general: it disabled Fable 5 and Mythos 5 even for some of Anthropic's employees, including those within the U.S.

Mitigation strategies

African institutions can take several steps to mitigate the risks associated with dependence on foreign AI technologies:

  • Local infrastructure development: Investing in local AI infrastructure can reduce dependence on foreign technologies and create economic opportunities. This requires significant investment in research and development, as well as a regulatory environment that encourages innovation and the growth of AI startups.
  • Diversification from U.S. providers: The Chinese initiative for global AI governance directly courts the Global South, and Beijing has turned that proposal into practice with a 2017 white paper. However, adopting Chinese AI could involve trade-offs in terms of security and privacy, as well as compatibility and interoperability issues with existing technologies. Moreover, China has a long history of human rights violations and domestic repression, which could raise concerns for African institutions seeking to collaborate with Chinese AI companies.
  • Development of local capabilities: Investing in the development of local AI capabilities can reduce dependence on foreign technologies and create economic opportunities. This requires significant investment in research and development, as well as a regulatory environment that encourages innovation and the growth of AI startups.
  • Regional collaboration: African institutions can work together to develop common standards and regulatory frameworks that promote the responsible and secure use of AI. This can also create opportunities for resource sharing and collaboration in research and development.

Implications and Future Perspectives

The episode of suspending access to Anthropic's AI models has highlighted a broader reality: technological dependence on foreign powers is a structural problem that requires long-term solutions. For African institutions, this means not only addressing the immediate consequences of a possible disruption of critical services but also preparing for a future in which AI plays an increasingly central role in economic and social development.

Investments in Local Infrastructure

Creating local AI infrastructure is a fundamental step, but it requires significant commitment from governments and the private sector. This includes building data centers, training qualified personnel, and developing policies that encourage technological innovation. However, it is essential that such infrastructures are not isolated but integrated into regional and global networks to maximize efficiency and effectiveness.

International Collaborations

Collaborations with international entities and non-governmental organizations can provide additional resources and technical expertise. For example, academic exchange programs and public-private partnerships can accelerate the development of local capabilities. Additionally, participation in global initiatives on AI governance can help African institutions influence international norms and standards.

Security and Digital Sovereignty

Data protection and cybersecurity are crucial aspects to ensure that local AI infrastructures are resilient against external threats. This includes adopting advanced cybersecurity measures and creating regulatory frameworks that protect digital sovereignty. The Malabo Convention represents a starting point, but it must be effectively implemented and adapted to emerging challenges.

Innovation and Development of New Models

Innovation should not be limited to the adoption of existing technologies but should include the development of new AI models tailored to local needs. This requires investments in research and development, as well as the creation of startup ecosystems that can compete globally. African institutions can leverage cultural diversity and the continent's unique challenges to create innovative AI solutions.

Education and Technical Skills

One of the biggest obstacles to the development of local capabilities is the lack of technical skills. Targeted educational programs, from primary school to university, can prepare future generations to work in the AI sector. Additionally, continuous training for existing professionals is essential to keep skills up-to-date.

Final Considerations

The challenge posed by technological dependence is complex but not insurmountable. With a combination of mitigation strategies, targeted investments, and international collaborations, African institutions can reduce risks and build a future where AI is a tool for sustainable development and innovation. The key to success lies in the ability to quickly adapt to technological changes and create an ecosystem conducive to innovation.

Editorial Note and Disclaimer

The guides and content published on GoYou are the result of independent research and analysis activities, for informational, educational, and in-depth purposes.

GoYou does not constitute a journalistic publication or an editorial product pursuant to Law No. 62/2001 and does not engage in real-time information activities.

The GoYou project does not provide professional, technical, legal, or financial advice and disclaims any liability for the improper use of the information published.

In the Crypto sector, every investment involves risks: readers are invited to always inform themselves autonomously before making any decision.