Artificial intelligence (AI) holds the promise of revolutionizing science and technology, offering solutions to grand challenges and accelerating scientific breakthroughs. However, as discussed by experts at the 48th annual AAAS Forum on Science and Technology Policy, the road ahead is fraught with challenges that require ongoing attention and action. The event, held on July 12 under the theme “Actionable Insights: Seeding Our Way Forward in AI,” brought together industry, academic, and government leaders to explore the transformative potential of AI and the necessary steps to navigate its complexities.
AI's recent advancements have begun to reshape various sectors, including the sciences. Tools powered by machine learning and generative AI are enabling researchers to tackle complex problems faster than ever before. Eric Horvitz, chief scientific officer at Microsoft, delivered the opening address, emphasizing that AI is still in its early stages of assisting scientists with modeling and reasoning. He highlighted how AI is poised to revolutionize the sciences by providing novel solutions to some of the world's most pressing challenges.
“AI is in the early days of helping scientists with modeling and reasoning,” Horvitz noted. He cited examples such as GitHub Copilot, a generative AI tool that has significantly increased programmers’ efficiency, and AI-driven literature reviews, which can synthesize and analyze vast amounts of research data in minutes. These advancements, he said, represent just the beginning of AI's potential to accelerate scientific discovery.
Horvitz also pointed to a collaboration between Microsoft and the Pacific Northwest National Laboratory, where AI was used to develop new lithium-based batteries. The AI-driven approach reduced the time and cost of the research dramatically, from an estimated 20 years and $2 billion to just 80 hours and $350,000.
While the benefits of AI are undeniable, the Forum’s speakers also urged caution, stressing the dual-use nature of technology. AI, they warned, could be used for harmful purposes, such as the development of chemical weapons or the spread of disinformation on an unprecedented scale. The risks associated with embedded AI—intelligent platforms integrated into everyday devices—were also discussed. Fay Cobb Payton, professor emerita at North Carolina State University, raised concerns about the potential for AI to erode human cognitive and computational skills.
Elham Tabassi, chief AI advisor at the National Institute for Standards and Technology, highlighted the need for AI systems to be trustworthy. “We want technology that's easy to do the right thing, difficult to do the wrong thing, and easy to recover if something unintentionally goes wrong or even intentionally,” she said. Tabassi emphasized the importance of transparency, accountability, and fairness in AI, along with the need for measurable standards to ensure trustworthiness.
The Forum concluded with a call to action, urging collaboration across government, academia, and industry to guide AI’s development responsibly. Speakers emphasized the importance of science-informed policymaking and the need for coordinated efforts to keep pace with AI advancements.
Sudip Parikh, CEO of AAAS, echoed this sentiment, encouraging attendees to take an active role in shaping the future of AI. “You can build the future,” he told the audience, emphasizing the collective responsibility of the scientific community to ensure AI is developed and used ethically and effectively.
Horvitz concluded on a hopeful note, envisioning the next 50 years as a time of unprecedented AI innovation and influence in the sciences. “We’re truly only at the foothills of an AI era in the sciences,” he said, underscoring the transformative potential of AI as long as its development is guided by careful and considered action.
Click here to read the original press release.
