From PGP to Mythos: a brief history of export controls that didn’t stop anyone
The US government recently restricted Anthropic from exporting its powerful AI models, Fable and Mythos, highlighting a new attempt to control advanced AI through export regulations. This event draws parallels with past, often unsuccessful, government efforts to control technologies like encryption and spyware.
The White House recently ordered Anthropic to restrict the export of its advanced AI models, Fable and Mythos, citing national security concerns. This move, which led Anthropic to pull both models, marks the first significant test of the U.S. government's ability to control frontier AI through export controls, a strategy previously applied with mixed results to encryption and spyware. The outcome of this standoff could significantly influence Anthropic's access to international markets and shape future regulations for other AI developers.
Anthropic had previously marketed Mythos as a critical tool for cybersecurity defense, limiting its access to approximately 150 vetted organizations due to its potent capabilities. The immediate cause for the ban reportedly stemmed from two incidents: Anthropic granting access to a South Korean telecom company suspected of ties to China, and Amazon researchers allegedly circumventing Fable 5's safeguards. These events prompted the Commerce Department to issue an export control directive, leading to Anthropic's rapid withdrawal of access to its products.
Historically, governments have struggled to effectively control the proliferation of what they deem dangerous cyber technologies. A notable example is the "Crypto Wars" of the 1990s, where the U.S. government attempted to restrict the export of encryption software like Pretty Good Privacy (PGP). Its creator, Phil Zimmermann, famously circumvented these controls by publishing PGP's source code, ultimately contributing to the widespread adoption of end-to-end encryption.
Similar challenges arose in the 2010s with Western-made spyware. Despite efforts to expand the Wassenaar Arrangement to regulate dual-use surveillance technologies, the agreement faced inherent weaknesses. Many countries, including major players in the spyware industry, did not adhere to it, and its application remained discretionary. This allowed companies to continue exporting surveillance tools to oppressive regimes, often by relocating operations to countries with more lenient export controls.
Related articles
We Added Too Many Guardrails and Broke Our Own Agent, Our AI VP of Finance Found a Setting We’d Missed for 8 Years, and an Agent Is Now the One Renewing Your Software: The Agents #007
This article discusses the complexities and unexpected breakthroughs encountered while deploying AI agents in a business setting. It highlights the critical balance in setting guardrails for AI, the diverging behaviors of agents across different platforms, and the surprising efficiency gains from integrating AI with existing financial tools.
Fika Jobs raises $4M to build a video-first hiring platform where AI agents interview candidates
Fika Jobs, a Stockholm-based startup, secured $4 million in pre-seed funding to advance its video-first hiring platform. This platform uses AI agents to conduct interviews and create short video profiles for job seekers, aiming to revolutionize the traditional recruitment process.
Business & StartupsHow to burst the AI bubble: Strike at its roots
Cory Doctorow
