National Security And Artificial Intelligence

The (U.S.) National Security Commission on Artificial Intelligence recently published its final report. The report is 756 pages long, so I am not going to claim that I’ve read it all. I read the introduction and some of the conclusion, and the chapter on microelectronics (basically, semiconductors and advanced packaging).

To give you a flavor, here are the opening paragraphs of the “Letter from the Chair and Vice-Chair”:

Americans have not yet grappled with just how profoundly the artificial intelligence (AI) revolution will impact our economy, national security, and welfare. Much remains to be learned about the power and limits of AI technologies. Nevertheless, big decisions need to be made now to accelerate AI innovation to benefit the United States and to defend against the malign uses of AI.
…
As a bipartisan commission of 15 technologists, national security professionals, business executives, and academic leaders, the National Security Commission on Artificial Intelligence (NSCAI) is delivering an uncomfortable message: America is not prepared to defend or compete in the AI era. This is the tough reality we must face. And it is this reality that demands comprehensive, whole-of-nation action. Our final report presents a strategy to defend against AI threats, responsibly employ AI for national security, and win the broader technology competition for the sake of our prosperity, security, and welfare. The U.S. government cannot do this alone. It needs committed partners in industry, academia, and civil society. And America needs to enlist its oldest allies and new partners to build a safer and freer world for the AI era.

The introduction to the report calls out four areas for special attention (I’ve edited the report here to keep the length of this post reasonable):

• Leadership: Ultimately, we have a duty to convince the leaders in the U.S. Government to make the hard decision and the down payment to win the AI era. In America, the buck stops with the President, and AI strategy starts in the White House.
• Talent: The human talent deficit is the government’s most conspicuous AI deficit and the single greatest inhibitor to buying, building, and fielding AI-enabled technologies for national security purposes. We need to build entirely new talent pipelines from scratch. We should establish a new Digital Service Academy and civilian National Reserve to grow tech talent with the same seriousness of purpose that we grow military officers. Just as important, the United States needs to win the international talent competition by improving both STEM education and our system for admitting and retaining highly skilled immigrants.
• Hardware: Microelectronics power all AI, and the United States no longer manufactures the world’s most sophisticated chips. We do not want to overstate the precariousness of our position, but given that the vast majority of cutting-edge chips are produced at a single plant separated by just 110 miles of water from our principal strategic competitor, we must reevaluate the meaning of supply chain resilience and security. The federal investment and incentives needed to revitalize domestic microchip fabrication—perhaps $35 billion— should be an easy decision when the alternative is relying on another country to produce the engines that power the machines that will shape the future.
• Innovation Investment: We worry that only a few big companies and powerful states will have the resources to make the biggest AI breakthroughs. The federal government must partner with U.S. companies to preserve American leadership and to support development of diverse AI applications that advance the national interest in the broadest sense. If anything, this report underplays the investments America will need to make. The $40 billion we recommend to expand and democratize federal AI research and development (R&D) is a modest down payment on future breakthroughs. We envision hundreds of billions in federal spending in the coming years.

The report is not talking about small amounts of money, as you can see from the last sentence above.

This is not a time for abstract criticism of industrial policy or fears of deficit spending to stand in the way of progress. In 1956, President Dwight Eisenhower, a fiscally conservative Republican, worked with a Democratic Congress to commit $10 billion to build the Interstate Highway System. That is $96 billion in today’s world. Surely we can make a similar investment in the nation’s future.

The chapter on microelectronics paints a similarly pessimistic outlook, although, for anyone who works in or around the semiconductor industry, the basic facts are not in dispute:

U.S. leadership in microelectronics is critical to overall U.S. leadership in artificial intelligence (AI). Several assessments underpin this argument:

• Hardware is a foundational element of the AI stack alongside data, algorithms, and talent.
• Exponential increases in computational power have driven the last decade of progress in machine learning (ML).
• After decades leading the microelectronics industry, the United States will soon source roughly 90% of all high-volume, leading-edge integrated-circuit production from countries in East Asia.
• This means the United States is almost entirely reliant on foreign sources for production of the cutting-edge semiconductors critical for defense systems and industry more broadly, leaving the U.S. supply chain vulnerable to disruption by foreign government action or natural disaster.
• Specialized hardware, novel packaging techniques such as heterogeneous integration and 3D stacking, and new types of devices will drive future AI developments as traditional architectures of silicon-based chipsets encounter diminishing marginal performance improvements.
• Demand for trusted microelectronics will only grow as the military and Intelligence Community (IC) continue to incorporate AI into mission-critical systems.

The chapter contains a lot of detailed recommendations, and these are reduced to actual recommended budgets in one of the appendices. I am not a Washington insider, to say the least, so I have no perspective on how likely any of this funding is to be allocated. Note that years in these funding recommendations are all government financial years, which run from October 1 to September 30. So 2022 means starting October 1, 2021, this year. Here are the recommendations just for microelectronics, the most relevant appendix to readers:

• Increase federal grants for microelectronics manufacturing: $15B ($3B per project on average)
• Increase funding for DARPA’s Electronics Resurgence Initiative (ERI): $400M in 2022 and $5B total ramping up as capacity to absorb the money increases
• Increase funding for National Science Foundation (NSF) semiconductor research: $300M in 2022 and $2.5B total 2022-2026 (ramping)
• Increase funding for Department of Energy semiconductor research: $400M for 2022 and $4.5B total for 2022-2026 (ramping)
• Establish the Advanced Packaging National Manufacturing Program: $1B for 2022, and $5B 2022-2026
• Establish the National Semiconductor Technology Center: $100M in 2022 and $2B total 2022-2026

There is one sense in which these amounts are a lot of money. On the other hand, when a state-of-the-art fab costs something like $15B, or a single (current-generation) EUV stepper $200M, they are inadequate. My guess is that developing a state-of-the-art process is also something measured in billions of dollars, if not tens of billions.

You can read the whole report at Final Report of the National Security Committee on Artificial Intelligence (reminder, it is a 750-page pdf).