Google has quietly updated the webpage for its Accountable AI and Human Centered Expertise (RAI-HCT) workforce, the workforce charged with conducting analysis into AI security, equity, and explainability, to clean mentions of “variety” and “fairness.”

A earlier model of the web page used language reminiscent of “marginalized communities,” “numerous,” “underrepresented teams,” and “fairness” to explain the RAI-HCT workforce’s work. That language has been eliminated, or in some circumstances changed with much less particular wording (e.g. “all,” “assorted,” and “quite a few” reasonably than “numerous”)

Google didn’t instantly reply to a request for remark.

Date: Feb 26 – March 6, 2025
Firm: @Google
Change: Scrubbed mentions of variety and fairness from the mission description of their Accountable AI workforce. pic.twitter.com/i9VvBcHMQ6

— The Midas Challenge Watchtower (@SafetyChanges) March 8, 2025

The adjustments, which have been noticed by watchdog group The Midas Challenge, come after Google deleted comparable language from its Startups Founders Fund grant web site. The corporate said in early February that it could get rid of its variety hiring targets and evaluation its variety, fairness, and inclusion (DEI) packages.

Google is among the many many large tech firms that have rolled back DEI initiatives because the Trump Administration targets what it characterizes as an “illegal” apply. Amazon and Meta have walked again DEI measures over the previous few months, and OpenAI just lately eliminated mentions of variety and inclusion from a webpage on its hiring practices. Apple, nevertheless, just lately pushed back against a shareholder proposal to finish its DEI packages.

Many of those firms, together with Google, have contracts with federal companies.