- This event has passed.
Assessing and Improving AI Trustworthiness: Current Contexts, Potential Paths
March 3 @ 5:00 pm - March 4 @ 10:00 pm GMT
The notion of AI trustworthiness, comprising a wide array of attributes such as robustness, accuracy, fairness, explainability, and privacy, presents a complicated set of challenges in assessing these aspects of trustworthiness. Accounting for the trustworthiness of a given AI system thus requires assessing a number of different performance measures of these key attributes and navigating a number of trade-offs among different aspects of trustworthiness. System designers, regulators, policymakers, and interested parties will each have to devise a set of tools to handle these challenges.
On March 3-4 from 12:00pm ET to 5:00pm ET, the National Academies of Sciences, Engineering, and Medicine (NASEM) will convene “Assessing and Improving AI Trustworthiness: Current Contexts, Potential Paths,” a public workshop sponsored by the National Institute of Standards and Technology to help think through this interrelated set of challenges. Informed by the panel discussions on current attempts at measuring and designing trustworthy systems, this workshop will work to produce initial ideas for activities and collaborations by academia, industry, and the public sector to improve the assessment of trustworthiness of AI systems, and recommendations for NIST and similar public bodies.