AI policy and the battle for computing power

Ben Buchanan discusses the evolution of AI as a revolutionary technology emerging from the private sector, contrasting it with past technologies that were largely government-funded. He notes that the U.S. government's reduced involvement in AI funding since the 1990s led to an "AI winter," while the resurgence of AI from 2012 has been driven by companies like Google and OpenAI. This shift poses challenges for U.S. policymakers, who now face a knowledge gap in AI development compared to previous technologies.

Buchanan emphasizes the importance of educating policymakers about AI, framing it as a geopolitical issue, especially in the context of U.S.-China relations. He argues that computing power, rather than data, is the key driver of technological advancement, with the U.S. and its allies holding a significant advantage in this area. He stresses that AI is fundamentally tied to physical computer chips, with Taiwan's TSMC being crucial for advanced semiconductor production. The geopolitical implications of controlling the computing supply chain are significant, particularly regarding China's access to advanced chips, which the Biden administration has sought to restrict.

The conversation highlights Taiwan's critical role in the global chip supply chain, with 97% of advanced chips produced there. The Biden administration has recognized this dependency as a national security vulnerability, prompting the bipartisan Chips and Science Act to boost U.S. chip manufacturing. Buchanan discusses the varying communication styles needed for different audiences, noting that policymakers prefer actionable information, while graduate students engage in deeper discussions about technology's implications.

He recalls bipartisan discussions during his government tenure, including a meeting with senators following Biden's executive order on AI. The speaker supports the decision to impose technology bans and restrictions, which received backing from many Republicans. However, he observes a significant policy shift between the first and second Trump administrations regarding the sale of advanced AI chips to China.

The speaker expresses concern about the ethical implications of AI in military contexts and emphasizes the importance of export controls on chip technology to prevent enhancing the Chinese military's capabilities. They assert that China will not produce chips as advanced as those in the U.S. until at least 2028. The speaker defends their administration's policy of prioritizing technology exports to democracies and American companies, referencing a National Security Memorandum aimed at fostering collaboration between government and the private sector.

The Department of Defense and intelligence community are committed to aligning AI technology with ethical standards, acknowledging concerns about AI's application in sensitive areas like combat. The speaker highlights the importance of autonomy in military applications, particularly in cyber operations and missile defense. Collaboration with other democracies is deemed essential for establishing norms and standards for military autonomy, culminating in a political declaration on autonomy principles agreed upon by 58 countries.

The ongoing debate between speed and caution in AI policy is underscored, especially as the private sector rapidly advances technology. Concerns about AI extend beyond military applications to safety-critical areas like transportation and the psychological effects of AI on social media. The speaker draws parallels between AI's transformative potential and the historical impact of the railroad in the late 1800s, emphasizing the need to balance speed and safety in technology.

Competition in technology, especially between the U.S. and China, poses risks of a "race to the bottom" that could compromise safety. The speaker emphasizes the importance of establishing a democratic lead for American companies to prioritize safety and trust in AI development. Geopolitical challenges complicate international coordination on AI policies, making it crucial to align international guidelines for AI safety.

The speaker acknowledges the State Department's efforts in international diplomacy but notes that deteriorating relationships due to tariffs and politics could harm businesses. They stress the need for dialogue with autocracies like China, highlighting the importance of maintaining engagement with both allies and adversaries on technology issues. Computing power is identified as a critical factor in national security, particularly concerning AI and cyber operations.

The speaker argues that the U.S. must take stronger actions regarding AI and cyber capabilities to maintain its advantage, especially in light of potential Chinese advancements if they gain access to U.S. technology. AI's role in national security is significant, impacting both offensive and defensive cyber operations. The conversation focuses on the intersection of AI policy and the battle for computing power, addressing the success of democracy in the age of AI through critical questions about technology invention, adoption, and alignment with democratic values. Concerns are raised regarding job impacts, the centralization of power, and the risks of a surveillance state, with a desire to assess progress in five to ten years based on these criteria.