A final rule again revising Sec. 1557 of the Affordable Care Act protects gender identity, sexual orientation and pregnancy from sex discrimination and brings back notice and other requirements that were dropped from the 2020 version of the rule.^[1] The HHS Office for Civil Rights (OCR) and CMS, which released the rule April 26, also require covered entities to have a Sec. 1557 coordinator and grievance procedure and give patients a “notice of availability” about free language assistance and auxiliary aids, among other things. And there’s something brand new at the intersection of discrimination protections and artificial intelligence (AI).

Covered entities are required “to make reasonable efforts” to identify patient care decision support tools used in their health programs and activities “that employ input variables or factors that measure race, color, national origin, sex, age, or disability” and then “make reasonable efforts to mitigate the risk of discrimination” resulting from use of the tools, according to the rule.

There’s no method for minimizing discrimination in the use of patient care decision support tools in the rule. “It establishes a general rule that covered entities must not discriminate on the basis of a protected class through the use of a wide range of tools and technologies, including AI,” said attorney Kyle Gotchy, with King & Spalding in Sacramento, California. The rule essentially requires covered entities not to let it happen to them, whether they developed the tools or are just the end users, he said.

But covered entities could be on the hook for penalties even if discrimination stemming from AI isn’t intentional. “It doesn’t matter if the AI tools don’t have a discriminatory animus to start with,” said attorney Rachel Carey, with Whiteford in Richmond, Virginia. “If certain tools result in discrimination in one of the covered areas, they can bring pretty big scrutiny.”