Today, I responded to the Department of Energy's Request for Information regarding how to mobilize talent for the Genesis Mission and develop an American workforce to advance AI for science and engineering.
*****
Thank you for the opportunity to submit recommendations for mobilizing talent for the Genesis Mission and Developing an American Workforce to Advance Artificial Intelligence (AI) for Science and Engineering. The Foundation for American Innovation is a 501(c)(3) organization that champions the technology, talent, and ideas essential to American prosperity, security, and flourishing. I am Ian Banks, the Director of Science Policy.
I write to address two components of the Request for Information: the place of promising secondary and primary school students and the failure of current STEM PhD programs to adequately prepare graduates for the demands of the innovation workforce of the coming years and decades. In addition to supporting undergraduate and graduate students, DOE should be considering promising high school students. For too long, federal policy has prioritized expanding equity in STEM over accelerating the training of the nation’s future top scientists and engineers. When those highly talented individuals do reach graduate school, they are trained along a singularly academic path and the interests of government and industry only rarely shape the activities and structure of STEM PhD programs. With the Genesis Mission, the DOE has the opportunity to drive the transformation of the scientific and technological ecosystem to support both of these efforts.
1. How can DOE catalyze research collaborations between DOE National Laboratories, universities, and industry to meet the goals of the Genesis program?
DOE should look beyond post-secondary institutions and begin identifying and cultivating exceptional STEM talent at the elementary and high school level. The Genesis Mission's goal of training 100,000 AI-capable scientists and engineers over the next decade will depend not only on reshaping graduate and undergraduate education but on ensuring the pipeline feeding into those programs is robust and inclusive.
There is strong precedent for this approach. In 1988, the National Science Foundation launched the Young Scholars Program, investing $3.7 million to create enrichment programs in science, engineering, and mathematics for talented high school students. Over its eight-year run, the program served roughly 18,000 students before being defunded in 1996 as a casualty of shifting federal priorities away from excellence-oriented programming. No comparable federal initiative has replaced it. Recently, promising philanthropic models have emerged to fill part of this gap. National Math Stars, founded in 2023, identifies mathematically extraordinary (1-1,000 talent based on math ability) students as early as second and third grade and provides each with mentorship, advanced coursework, summer STEM experiences, and family support over a ten-year period through high school graduation. This demonstrates the feasibility and demand for systematic talent identification and long-term development starting as early as the elementary school level.
Longitudinal research tracking of mathematically precocious students over four decades has shown that early identification and acceleration of exceptional young people produces outsized returns: cohorts identified before age 13 went on to earn tenure at major research universities, secure hundreds of patents, and generate hundreds of millions in research funding. Meta-analyses of academic acceleration interventions have shown that students who receive advanced coursework outperform peers on standardized achievement, college attainment, and career outcomes without negative social or emotional effects.
Additionally, Harvard's “Lost Einsteins” research demonstrates that the nation is systematically failing to identify high-potential students and that closing this gap could quadruple the U.S. rate of innovation. A DOE Young Scholars program would connect promising high school students directly with National Laboratory scientists and engineers through summer residencies, mentorship, and project-based learning tied to Genesis Mission challenge areas such as fusion energy, advanced materials, and AI-driven experimental science. This program should be designed to integrate with existing National Laboratory educational programming and user facility access, so that Young Scholars participate in authentic research environments rather than standalone enrichment activities. National Labs already host visiting researchers, graduate students, and postdoctoral fellows; extending structured access to exceptional high school students would add a critical early rung to that existing ladder. DOE could also partner with organizations like National Math Stars to identify and recruit students who have already been flagged as exceptional through rigorous, equity-focused screening processes, creating a seamless handoff from elementary and middle school talent development into DOE's high school and eventually post-secondary pipeline.
This would not only expand and diversify the long-term talent pipeline for DOE's workforce needs but would give the United States a structured answer to the talent development strategies of its geopolitical competitors. China currently selects roughly 100,000 teenagers annually for advanced science instruction as part of a multi-decade strategy that has already paid dividends in AI and other frontier technologies.
2. How can DOE incentivize partnerships between universities, DOE National Laboratories, industry, and philanthropic organizations to 1) establish new training paths for bachelor’s and master’s degrees focused on dual competencies in AI and scientific/engineering disciplines and 2) provide innovative experiences to prepare doctoral students and post-doctoral associates for careers in AI for science?
The current American STEM PhD is too slow and too narrowly focused to meet the urgency of the Genesis Mission. The average U.S. STEM doctorate takes nearly six years to complete — compared to four years or less in Germany and France — and three-quarters of STEM PhD graduates ultimately leave academia for careers outside of research universities, suggesting that the system is training students for a destination most of them will never reach. As Prineha Narang and I recently argued in Science magazine, “Reimagining the STEM PhD,” if one goal of advanced training is to produce professionals who can drive innovation and apply new ideas to create value, then programs must be designed to support that outcome and not just prepare students for careers in curiosity-driven academic research. China now accounts for 56 percent of publications in elite science journals compared to just 10 percent for the United States, a dramatic reversal from five years earlier that demands a remodeling of U.S. graduate training.
DOE is uniquely positioned to shape this remodeling. The Department should cultivate partnerships between universities, National Laboratories, industry, and philanthropic organizations to establish a STEM Innovation PhD fellowship program aligned with Genesis Mission priorities and modeled on the very successful NSF Graduate Research Fellowship. This track would compress time-to-degree through structured, project-based research from day one looping in cutting edge work from the National Labs, industry AI deployments, and autonomous laboratory environments. Funding could follow a hybrid model combining federal support with industry contributions for specialized equipment, internships, and access to frontier AI hardware and software platforms. DOE’s fellowship program would demonstrate that the United States is serious about building this pipeline from the ground up, and that the exceptional students who are admitted to this track can expect federal backing.
Critically, DOE should also use these partnerships to extend the pipeline downward. The same university-lab-industry-philanthropy consortia that design doctoral and master's programs could also provide access to summer enrichment programs for exceptional high school students who have the potential to become the next generation of builders, hackers, and scholars in science, engineering, and technology. Connecting these students to leading research environments before they even begin undergraduate study would accelerate their development and deepen the domestic talent pool feeding into dual-competency degree programs.
4. Beyond funding, what other opportunities could DOE, including its National Laboratories and user facilities, bring to these partnerships?
Beyond funding, DOE's most distinctive asset is the National Laboratory system itself. DOE should draw on this unmatched resource to reach the next generation of scientists and engineers before they arrive at college. The Young Scholars program, suggested above, can place exceptional high school students in structured summer residencies at National Laboratories, where they would work alongside scientists and engineers on authentic research tied to Genesis Mission challenge areas. Unlike generic STEM summer camps, these residencies would immerse students in the scientific workflows at the frontier DOE is trying to explore. The program could pair students with a laboratory mentor who provides technical guidance and sustained contact through the academic year, bridging the transition from high school to undergraduate study and drawing students naturally into the dual-competency pipeline.
The reach of this effort should extend to the educators and counselors who first encounter these students. DOE and its National Laboratories should offer training and resources to high school teachers and counselors to help them identify and support high-potential STEM students in their schools. National Labs could also help connect teachers to laboratory scientists as content advisors; and create referral pathways so that students identified through universal screening or programs like National Math Stars have a clear route into DOE's summer programming and, eventually, into the broader Genesis Mission workforce pipeline.
