On March 14, Sens. Mike Rounds (R-SD) and Martin Heinrich’s (D-NM) announced the American Science Acceleration Project (ASAP), a bipartisan initiative to make U.S. science ten times faster by 2030. This ambitious effort couldn’t come at a more critical moment. American innovation is losing steam, and without urgent action, we risk falling behind in the global technology race.
The mid-20th century was a golden age for American science. From cracking the code of DNA to landing on the Moon, the U.S. led the world in breakthrough discoveries. But today, despite unparalleled resources and talent, true game-changing innovations are becoming rarer. Various measures suggest a scientific slowdown is happening. Productivity growth—a key indicator of scientific and technological progress—has slowed from 3.2 percent per year in the 1950s to just 1.8 percent today.
Meanwhile, China is rapidly closing the innovation gap. Beijing now leads in critical strategic technologies like electric batteries, hypersonics and advanced radio-frequency communications. It’s no wonder that three-fourths of Americans believe today’s children will be worse off than their parents.
But this slowdown isn't inevitable. A new partnership between American scientists and artificial intelligence (AI) tools can usher in a new era of scientific discovery and economic prosperity. AI tools are already helping researchers amplify their capabilities. Last year's Nobel Prize in Chemistry recognized AlphaFold, which demonstrates how human-AI collaboration can solve longstanding scientific challenges like protein structure prediction. When thoughtfully integrated into researcher workflows, AI helps scientists generate hypotheses, design experiments and interpret large datasets, accelerating the pace at which human creativity and expertise can drive breakthrough discoveries.
These innovations were only made possible through existing scientific infrastructure. AlphaFold, for example, was trained on public, often open-source datasets like the Protein Data Bank. The next generation of breakthrough-enabling AI will need new data platforms that are interoperable, easy to use and secure. There also needs to be extensive buildout of advanced compute infrastructure – new data centers, ultra-high-speed networks and energy infrastructure. The Trump Administration’s plans to expedite power plant construction for AI mark an important first step, but more is needed.
To make this vision a reality, the U.S. government and private industry must step up with strategic investments. We must build a "superhighway for science" that connects scientific data, advanced computing, and AI tools to help America reclaim the scientific and technological frontier. America’s scientific talent should have the tools they need to solve today's grand challenges, from curing cancer to creating new materials that could revolutionize everything from batteries to buildings.
This isn’t about creating a new bureaucracy. Rather,it’s a call to action for the government, civil society, industry, and academia to create new partnerships and build the scientific infrastructure of the future. With the cuts and financial constraints of today, we need efficient ways to maximize the impact of every dollar.
This next-generation scientific infrastructure must be built with security at its core. As America's scientific capabilities expand, so too will attempts by adversaries to compromise these systems or steal valuable IP, especially AI system components such as model weights or innovations with national security relevance like biotech or quantum computing. The new State Department Bureau of Emerging Threats is an encouraging sign that the Administration recognizes the importance of these risks on the global stage. Given the dual-use potential of advanced AI systems, we need safeguards to prevent misuse, such as dangerous capability evaluations and access controls. Differential access arrangements would allow researchers to use advanced science models to ensure safety and security, such as developing medical countermeasures or strengthening cybersecurity, while minimizing proliferation risks.
Equally important is the fundamental work needed to make AI systems more transparent and trustworthy. Recent polling shows Americans have limited trust in AI, which could slow uptake of impactful new tools. For AI to have a transformative impact on science, these tools must be able to explain their reasoning, demonstrate reliability and integrate seamlessly into existing research workflows. Scientists need to understand not just what an AI system predicts, but why it makes those predictions and how confident they should be in its conclusions.
The path forward isn't through vague promises of an artificial general intelligence-led future. It's through a focused national mission to supercharge scientific discovery. This isn't just about building powerful AI systems; it's about putting revolutionary tools in the hands of America's scientists to solve real-world challenges like curing disease. By bringing together government, industry, the scientific community and civic leadership, America can return to its place as the world's engine of discovery.
When Neil Armstrong walked on the Moon, Americans knew the country could achieve anything. Now, it's time to rekindle that spirit of possibility and launch a new golden age of American innovation.
Jam Kraprayoon is a researcher at the Institute for AI Policy and Strategy and Fellow at the Federation of American Scientists.
Read Full Article »
