A new report from the President’s Council of Advisors on Science and Technology (PCAST), which is set to be released next week, focuses on several strong use cases for artificial intelligence (AI) tools in scientific research.
PCAST – which unanimously voted on April 23 to send the report to President Biden’s desk – was tasked with exploring the synergy between AI and science in the administration’s AI executive order issued last October.
“Supercharging Research: Harnessing Artificial Intelligence to Meet Global Challenges” explores how AI is already transforming science and will continue to do so. It also offers recommendations for how the Federal government can support AI-assisted research and development.
Terence Tao, co-chair of the PCAST Working Group on AI and professor at UCLA, said the report finds that AI is starting to remove many of the barriers that make scientific research slow and expensive.
Tao said that when the report is published next week, it will include several examples of how AI will transform science, but during his Tuesday presentation he highlighted three of those: identify candidate solutions to scientific problems; accelerate and enhance scientific simulations and models; and analyze new types of data.
Leveraging AI to identity solutions in therapeutic drugs is one prominent example. Tao said that currently, AI is already being leveraged to do things like design vaccines, but in the future it could be used to help treat cancer.
The report highlights that AI can also be used to further accelerate modeling for climate science. Currently, AI is being used for weather forecasting but in the future, whole Earth AI models will help with things like catastrophe modeling.
In his presentation, Tao also highlighted that AI can be used for new types of data in health and wellness. AI is currently being leveraged to aid physicians in making early diagnoses, but the technology could be used in the future to provide ultra-personalized medicine and healthcare.
PCAST’s report highlights its vision of AI-assisted research and development: “To gain the greatest net benefit from AI adoption in science, we envisage pursuing three broad goals: empowerment of human scientists; responsible use of AI tools; and sharing of AI resources.”
The report offers five recommendations to advance these three goals:
- Support both basic and applied research in AI that involves collaborations across academia, industry, national laboratories, and Federal agencies;
- Encourage innovative approaches to integrating AI assistance into scientific workflows;
- Adopt principles of responsible, transparent, and trustworthy AI use throughout all stages of the scientific research process;
- Expand existing efforts to broadly and equitably share basic AI resources; and
- Expand secure access to Federal data sets for approved critical research needs, with appropriate protections and safeguards.
Many PCAST members specifically highlighted the National Science Foundation’s National AI Research Resource (NAIRR) pilot, which was launched early this year, as a prime example of how the Federal government can meet the three broad goals of using AI in science laid out in the new report.
The PCAST report calls for full funding of the NAIRR – like many Federal leaders and legislators have recently.
