What and Why: AI at Los Alamos

Download a print-friendly version of this article.

Seemingly overnight, AI is producing podcasts, driving cars, and influencing the global economy. But it could still be better at science. At Los Alamos National Laboratory, our scientists are using and developing AI that can revolutionize science and accelerate discovery. By harnessing our unparalleled computational resources, experimental capabilities, and world-class talent, we’re opening new scientific frontiers and reducing threats against our country and the world.

While private companies lead the world in AI development, Los Alamos has complementary goals for our science and national security missions. Therefore, we are partnering with companies to accelerate the application of large-scale AI to national challenges. For example, mainstream AI models can’t be used for sensitive scientific data, but specially adapted versions can. At Los Alamos, our scientists are evaluating AI's security risks while designing and building powerful AI that promises to deliver a new golden age of scientific discovery. Here's how AI is making our country and world safer. 

Computational Breakthroughs

Scientific research in complex fields, like nuclear physics and materials science, has always been limited by the processing and data-handling speeds of traditional computational methods. At Los Alamos, AI is transforming our approach to scientific computing. Our data scientists are using advanced AI to uncover hidden patterns— strengthening our nuclear deterrent, modernizing our electrical grids, and ushering in the energy systems of the future. AI is being integrated into nearly every aspect of our work, amplifying the impact of our scientific computers. 

Next-generation computers could match the power (and size) of the human brain. More about AI in Computation.

Integrated Deterrence

The global security landscape has never been more dynamic, with new threats emerging across the globe and even in space. Los Alamos scientists leverage AI's predictive analytics and simulation abilities to stay ahead of emerging threats. Through the rapid processing of diverse and wide-reaching datasets, AI helps us monitor the global nuclear landscape, ensure our own stockpile is modern and effective, develop strategies for nuclear deterrence in space, and even design ways to detect when AI is being used for harm.

Can scientists trust AI to solve high-stakes scientific questions? More about AI in Deterrence.

Biosecurity and Emerging Threats Preparedness

In the modern world, humanity is more connected than ever before, enabling life-threatening pathogens to spread rapidly, endangering health and economies. One key to combating emerging diseases is scientific data. At Los Alamos, our scientists are combining complex data with AI to predict outbreaks, anticipate how pathogens will spread, and develop detection schemes, drugs, and vaccines for infectious diseases.  

Advanced AI could speed the design of new drugs and vaccines. More about AI in Biosecurity.

“It looks very likely that AI is going to completely change how we do our work.”
—Pat Fitch, deputy Laboratory director for Science, Technology, and Engineering

Experimental Advances

Los Alamos maintains and develops the world's most sophisticated defense systems without ever conducting full-scale nuclear tests—a job called science-based stewardship. This critical work relies on computer models that meticulously replicate every aspect of a weapon’s performance. But these simulations hinge on data gathered from real-world experiments conducted at highly specialized particle accelerators: the Los Alamos Neutron Science Center and the Dual-Axis Radiographic Hydrodynamic Test facility. At these Los Alamos facilities, scientists ask precise physics questions that can be answered nowhere else. By integrating AI into the pipeline of nuclear data production, our scientists can design targeted experiments and analyze data faster, ensuring our nation’s stockpile is safe, reliable, and the most advanced in the world.  

Cutting through years of iteration to get quality data and reliable results faster—with AI. More about AI in Experimentation.

National Security

Simulating real-world applications of high-stakes science, like experimental energy sources or nuclear proliferation, requires stitching together millions or even billions of data points. The task demands significant resources and involves trial and error. AI can rapidly analyze vast amounts of data, quickly providing answers to an enormous range of complex questions. This AI-driven approach enables faster, more informed decision making in critical areas of national security and energy dominance.  

AI is modernizing Los Alamos’ computer languages. Can old codes on new computers reveal fresh scientific insights? More about AI in Modeling.

“AI is a huge part of our future, and we have to figure out how to keep the Lab moving as fast as possible.”
—Chris Wright, Secretary of Energy

Leaders in AI

Earl Lawrence is the statistician leading a major Los Alamos effort to deploy AI for science. Over a multiyear project, Lawrence and a cross-disciplinary team of more than 100 of the Laboratory’s top data scientists are developing and applying AI models that can solve tangible scientific problems with access to specialized datasets. He’s also developing AI to design new experiments, automate work in the lab, and control manufacturing.

Can a single AI model advance any field of science? More on AI for Science.

“If we succeed, we could rapidly advance the pace of science on topics that benefit all of society. There are also scary parts of this technology that we at Los Alamos understand better than most.”
—Earl Lawrence, head of the Laboratory’s Artificial Intelligence for Mission project

Jason Pruet, the director of the Laboratory’s National Security AI Office, is building partnerships with universities and tech companies to use existing AI models to solve complex national security challenges.

Inside the Lab’s race to harness AI’s vast potential—and mitigate its potential harms. More on AI for National Security.

“AI is starting to feel like the next great foundation for scientific progress. The government must lead the work of harnessing this technology for the public good.”
—Jason Pruet, Director of the Laboratory’s National Security AI Office