Preparing students with the skills needed for an AI-shaped future: A Q&A with Jim Short and Angela DeBarger
Schools have a powerful opportunity and responsibility to shape new technologies to strengthen our democracy and our economy. Students need to build the competencies that matter most, from critical thinking to evidence-based reasoning and collaboration, and actively practice and strengthen their agency as leaders in their communities. Science classrooms offer one of the most powerful levers for developing these skills and practices at scale, yet less than a quarter of high school seniors are proficient in science, and the infrastructure to change that barely exists.
In this conversation, Kumar Garg talks with Jim Short and Angela DeBarger, who are incubating a thesis-driven, philanthropic fund to instill durable skills through science learning and teaching, preparing students for long-term economic mobility and democratic participation.
Kumar Garg: How big is this problem, really?
Jim Short: The data are stark. Only 22% of 12th graders and 31% of 8th graders demonstrate proficiency in science, at the very moment STEM jobs are projected to grow more than three times as fast as other occupations. The U.S. is on track for a shortage of 1.4 million STEM jobs for computer scientists, engineers, and technicians by 2030. But the deeper problem is, in a world defined by artificial intelligence, climate change, and pervasive misinformation, every citizen needs to be able to evaluate evidence and reason from it. Scientific literacy is a prerequisite for democratic participation, not just economic mobility.
Angela DeBarger: We need to prepare students to think creatively and scientifically, and this needs to begin from the earliest ages. Students come to school excited to learn about how the world works, but elementary schools have reduced science instructional time to less than 20 minutes per week. Nationally, reductions in science education funding limit opportunities for supporting innovative teaching and building strong local implementation systems. The problem isn’t that we don’t know what good science education looks like. It’s that systems are not designed to deliver it consistently and equitably. That’s the structural gap the Catalyst Fund is being designed to fill.
KG: If we know what good science education looks like, why aren’t more students experiencing it?
JS: Students are experiencing better science instruction when teachers use high-quality curriculum, like OpenSciEd. OpenSciEd is a freely available, research-based, open-source K-12 science curriculum that is being used to teach students in 1,270 school districts across all 50 states. It works, and there is evidence to show the impact on students: across a 10-state field test with more than 10,000 middle school students, it showed equitable outcomes across race, gender, and language background. Students from every demographic reported that their ideas mattered and that they felt like genuine participants in building scientific knowledge. We have done a deep analysis of six states where more than 330,000 middle school students are in OpenSciEd classrooms, but sustained implementation support is simply not there.
AD: Access to high-quality instructional materials is necessary, but it’s not sufficient. What’s missing is the social and technical infrastructure to implement it well. As opposed to one-off workshops, teachers need sustained professional learning anchored in the curriculum, where they engage in the same kind of inquiry and reflection that they expect from their students. School leaders who support teachers also need tools to diagnose implementation challenges and act on them. None of that infrastructure has been built at scale.
KG: Why is this the moment to act?
JS: Three things are converging right now. First, validated open-source instructional materials exist and are already being used to teach science in districts across every state, which means we’re not starting from scratch. Second, states are showing unprecedented commitment to what they call “Portrait of a Graduate” outcomes: durable, cross-cutting competencies that go beyond test scores. State and business leaders are aligned on what they want students to be able to do, and this is an agenda-setting moment with governors’ races in 39 states this November. Science education is one of the clearest practical pathways to those goals. Third, and this is what makes the timing genuinely urgent, the use of AI to support instruction. The same technology that is disrupting the labor market is also, if used responsibly, a transformative tool for improving instruction.
AD: When it comes to AI, we’re developing applications that strengthen opportunities for sensemaking and collaboration for both educators and students. We can now build AI tools that give teachers real-time feedback, support students in collaborative reasoning, and give district leaders actionable implementation data. That capability didn’t exist five years ago. The window to shape how AI enters science classrooms to strengthen human reasoning and connection is open now.
KG: How does this work actually turn opportunity into a scalable system?
AD: The Catalyst Fund is being designed as an initiative focused on two mutually reinforcing levers that advance future-ready skills systemically. The first is AI-enabled tools. We’re envisioning a Collaborative Learning Workspace that supports student groups in evidence-based sensemaking; a Science Teaching Tool that functions as a virtual coach for teachers who are new to high-quality instructional materials; and an Implementation Navigation Tool that helps state and district leaders interpret implementation data and generate targeted change strategies. All three will be built on open-source infrastructure, grounded in research, and designed to strengthen student reasoning and teacher judgment.
JS: The second lever is local capacity and coherent implementation infrastructure. We’re targeting three states initially, using a hub-and-spoke model where an anchor district with OpenSciEd experience extends support to surrounding districts, amplified by regional service agencies. Our two national implementation partners, BSCS Science Learning for curriculum-based professional learning, and Columbia University’s Center for Public Research and Leadership for change management, ensure the model is rigorous and replicable. These two levers are designed to work together: AI tools are most powerful inside strong local systems, and local capacity is amplified by timely, curriculum-aligned data. By the end of three years, we project 200,000+ middle school students experiencing authentic, inquiry-based science learning; 1,500+ teachers effectively implementing high-quality curriculum with AI-enabled support; and three state hubs generating the implementation evidence and playbooks that other states can replicate.
KG: What drew each of you personally to this work?
JS: I’ve spent over three decades in science education, as a high school biology teacher, a district science coordinator in Denver, the founding director of a center for science teaching and learning at the American Museum of Natural History, and a program director at Carnegie Corporation of New York where we launched OpenSciEd as a $40 million initiative. I’ve seen, over and over again, that when students get the chance to actually do science, to ask real questions, wrestle with evidence, build explanations together, something clicks. They become critical thinkers who can apply science knowledge to explain natural phenomena and design solutions to problems.That experience is transformative, and often not accessible to the students who need it most. What drives me is the belief that this gap is not inevitable. We have the tools we need to close it.
AD: For me, it starts with a conviction that school needs to be a place where students find joy, connection, and meaning in learning. Early in my research career, I realized impactful and long-lasting change in classrooms and school systems is only possible through partnerships among practitioners, researchers, community members, and students. The Catalyst Fund is, at its core, designed to expand access to meaningful, connected learning that will inspire and prepare the next generation of leaders.
KG: What should funders and partners do right now?
AD: The Catalyst Fund has an immediate philanthropic entry point that is both concrete and catalytic. A $1.5 million scoping grant from the William and Flora Hewlett Foundation is already funding the design phase involving strategy refinement, state and district engagement, convenings with partners, and AI tool prototype development and testing. Additional funding can unlock the first phase of implementation: activating state hub contracts, AI tool development expansion, and enrolling the first cohort of teachers in curriculum-based professional learning.
JS: Funding partners can also have real strategic influence, including ongoing dialogue with fund leadership on state selection, AI tool development priorities, and implementation learning as it unfolds. It’s an opportunity to help shape a model that becomes a blueprint for how states build science education infrastructure in an AI-shaped world. The urgency is real, the infrastructure is ready to build, and the window to act is now.