Microsoft Elevates Responsible AI with New Trusted Technology Lead

In February 2026 Microsoft named accessibility veteran Jenny Lay-Flurrie head of the Trusted Technology Group (TTG). TTG unifies responsible-technology programs—accessibility, fairness, transparency, inclusiveness and accountability—under a single top-down structure, tasked with building AI that is correct from inception and remains reliable.

CNBC

Microsoft’s responsible-tech agenda dates to Bill Gates’s 2002 “Trustworthy Computing” memo, which prioritized reliability over rapid feature rollout. The agenda gained new urgency after the Trump administration’s March 20, 2026 national AI legislative framework, which framed AI leadership as a competitive race. TTG was created to embed ethical safeguards early in product cycles.

CNBC

To improve blind-user support, Microsoft bought more than 20 million minutes of multimodal video from Be My Eyes, a nonprofit that connects blind volunteers with live assistance. The footage—people navigating with canes, dogs and locating objects—was anonymized, blurring faces, then used to train models that previously generated images of blind people wearing full-on blindfolds.

CNBC

• 20 million minutes of multimodal data from Be My Eyes.

CNBC

• About 15,000 Microsoft jobs cut in 2025 across sales, gaming and customer-facing roles, with new hires focused on AI infrastructure.

CNBC

• The disability employee group received early access to Microsoft Copilot, the first community rollout.

CNBC

Microsoft says accountability is essential for AI regardless of outcomes, a principle reflected in TTG’s rapid feedback loop that collects, iterates, tests and resolves user concerns as quickly as possible. The firm also offers free Microsoft Learn modules on responsible AI to students, academics and developers, positioning internal lessons as public resources.

CNBC

Annie Brown, CEO of bias-mitigation firm Reliabl, warned that diverse data alone does not eliminate bias. She noted that insufficient attention to metadata labeling can embed bias, urging Microsoft to scrutinize how training images are tagged as well as the data volume.

CNBC

• “How do we make sure that we build it right? And how can we make sure that it stays right?” — Jenny Lay-Flurrie, Head of Trusted Technology Group, Microsoft

CNBC

• “Some of the generated imagery of blind people came back with people wearing these horrible full-on blindfolds,” — Jenny Lay-Flurrie, Microsoft

CNBC

• “More diverse data is just part of it," said Brown.” — Annie Brown, CEO, Reliabl

CNBC

• “people should be accountable for AI” — Microsoft spokesperson

CNBC

The source material agrees on the 20 million minutes of data and the 15,000 job cuts, but no independent evaluation of the accessibility improvements has been published, leaving their real-world impact unverified.

CNBC

Microsoft plans to broaden its data-collection partnerships and to embed disabled individuals throughout AI development, not as charitable participants but as decision-makers. Updated Microsoft Learn modules aim to disseminate TTG’s responsible-AI practices industry-wide, while internal metrics will track the effectiveness of newly introduced accessibility features. Disability entrepreneur Diego Mariscal, CEO and founder of 2Gether-International, called the new role a testament to inclusive design.

CNBC