Standing Committee on Use of AI in Science: Integrity, Trust, and Innovation in the Scientific Enterprise
The National Academies will convene a standing committee that will explore how artificial intelligence (AI) is challenging the norms, standards, and governance used to ensure research trustworthiness, reproducability, and alignment with societal values. The standing committee will host a series of public sessions that will be organized around key topics, including research integrity and reproducibility; transparency and trustworthiness; authorship and accountability; and peer review.
In formation
Statement of Task
The National Academies will convene a standing committee that will explore how artificial intelligence (AI) is challenging the norms, standards, and governance used to ensure research trustworthiness, reproducibility, and alignment with societal values. The standing committee will host a series of public sessions that will be organized around key topics, including research integrity and reproducibility; transparency and trustworthiness; authorship and accountability; and peer review.
A small group of experts selected by the National Academies will author issue papers that will draw in part from public session discussions to identify challenges and opportunities for aligning AI use with the norms of science, highlight best practices, and identify implications for an effective, fair, and future-ready scientific enterprise across three areas: integrity, transparency, and reproducibility; accountability and authorship; and peer review and opens science. The collection of issue papers will be written for decisionmakers and constituent groups, including academia, industry, publishers, funders and sponsors of research, government officials at all levels, and the broader public. Specific questions of interest include:
- How is AI reshaping the methods and epistemology of scientific inquiry?
- How does our understanding of reproducibility and validation change with different uses of AI?
- How can authorship, credit, and accountability be defined assessed/evaluated when AI systems contribute to scientific outputs?
- What levels and types of transparency and disclosure are necessary for AI-assisted research and peer review?
- How can open science principles be adapted to AI models? What are the implications for interoperable research communication to ensure reliable AI-enabled science?
The standing committee’s activities will seek to characterize the evolving use of AI in science and emerging frameworks, governance approaches, and practices (including from the international experience) that can enable AI to accelerate discovery while strengthening integrity, transparency, and public trust.
Contributors
Sponsors
Internal Funding
Staff
Karla Hagan
Lead
Jon Eisenberg
Lead
Emanuel Robinson
Lead
Tho Nguyen
Lead
Shenae Bradley