NIST artificial intelligence work aims to create taxonomy for establishing trust

The National Institute of Standards and Technology is developing ways to explain how to address risk in artificial intelligence as a means of increasing trust.

To make AI more trustworthy, there needs to be an understanding of “building blocks” or technical requirements first, NIST’s Elham Tabassi said Thursday at a NIST advisory board meeting.

Tabassi said there are “aspirational” principles that need to be translated into technical requirements and are “clear, implementable and testable,” which will define “what constitutes trust...