Donald Trump's AI initiative, "The Genesis Mission," includes tasks related to the U.S. nuclear weapons complex. Journalists from Scientific American highlighted this after the Department of Energy released a list of its focus areas.
The initiative is presented as a program to accelerate scientific discoveries using artificial intelligence. However, among the first 26 tasks, there is a section related to nuclear deterrence, weapons production, non-proliferation, and nuclear forensics.
What is the "Genesis Mission"?
On November 24, 2025, Trump signed an executive order launching new federal research and development programs in artificial intelligence. The U.S. Department of Energy is tasked with deploying the "American Science and Security Platform"—an infrastructure to integrate supercomputers from national laboratories, secure cloud AI environments, scientific datasets, specialized models, and tools for AI experimentation.
The executive order describes the initiative as a means to accelerate scientific discoveries, enhance national security, and ensure "energy dominance." The platform is expected to support model training, simulations, AI agents, workflow automation, and the use of federal, academic, and private datasets. The Department of Energy is responsible for implementation, with Deputy Secretary for Science Dario Gil leading the project.
According to the timeline, the department was to identify at least 20 national science and technology tasks within 60 days, conduct an inventory of computing resources within 90 days, select initial datasets and models within 120 days, and demonstrate initial platform functionality for at least one task within 270 days.
Nuclear Tasks in the List
On February 12, the Department of Energy published a list of 26 tasks for the "Genesis Mission." These tasks cover energy, science, manufacturing, critical materials, quantum technologies, data centers, and national security.
Scientific American compared public statements from officials with this list and identified seven areas related to the nuclear complex and national security. The published documents do not indicate that AI will be given the authority to make decisions regarding the use of nuclear weapons.
- Assessment of nuclear threats, readiness, and response. The Department of Energy proposes creating a multimodal AI system that integrates sensor data, modeling, and intelligence reports. The goal is to reduce the time from threat detection to response from days to hours.
- Digitization of historical nuclear data. This involves archiving classified nuclear weapons experiments and tests, as well as unclassified materials on nuclear science spanning over eight decades. Plans include using text recognition, computer vision, information extraction, and 3D reconstruction to convert paper reports, photographs, and blueprints into searchable and model-ready datasets.
- Increasing throughput at research nuclear facilities. The department describes an AI "operating system for facilities" that will assist in planning experiments, adjusting them in real-time, and integrating diagnostic data with modeling.
- Integration of design and manufacturing for nuclear deterrence. The description mentions reducing the cycle between design and manufacturing units to deliver new weapon systems "in a fraction of the traditional time and cost."
- Non-proliferation. The Department of Energy aims to use multimodal models and analytical agents for joint analysis of satellite imagery, sensor data, open-source information, and government data. The system should identify anomalies related to potential proliferation of nuclear materials.
- Optimization of operations at high-risk facilities. The department proposes using verifiable AI systems based on large language models and agents to automate safety requirement analysis, documentation preparation, and work planning. The goal is to reduce planning and documentation time by more than 50%.
- Nuclear forensics. The Department of Energy plans to apply multimodal AI for accelerated analysis of samples and debris, determining their origin, and reconstructing the likely history of the material following an incident.
There is also a separate task focused on accelerating the development, production, and certification of materials for strategic deterrence, mentioning AI agents for discovering and testing new plutonium purification methods.
AI Does Not Get the "Button"
Scientific American emphasized that the discussed scenarios do not imply transferring decision-making authority regarding a theoretical nuclear strike to AI. The publication cites Herbert Lin, a senior research scholar at Stanford's Center for International Security and Cooperation, stating that no one in the military, National Security Council, or procurement system is proposing to hand over launch authority to an autonomous system.
In February 2026, a Pentagon representative also told The Washington Post that the decision to use nuclear weapons remains with humans, and the transfer of such a function to AI is not being considered. Experts express concerns about the broader implementation of AI in threat analysis, production, planning, experimentation, and risk assessment. Lin warned that AI should not be viewed as a source of guaranteed truth, especially when assessing real or potential attacks.
Bahrad Sokhansanj, a researcher at the Institute for Law and AI, noted that the Department of Energy is well-suited for such a program due to its existing national laboratories, supercomputers, and data. However, he also pointed out the management risks: accelerating science could impact not only beneficial but also potentially dangerous areas, including biological and defense research.
It was previously reported that U.S. national laboratories would apply new OpenAI models for scientific research and nuclear weapons security.
In May, the Pentagon entered into agreements with Nvidia, Microsoft, Reflection, and Amazon Web Services to use AI tools in classified military environments.
Subsequently, the U.S. military convened leading contractors for deeper integration of artificial intelligence into military systems while maintaining human oversight.