Microsoft’s new responsible tech lead on high-speed AI development

Completely responsible, reliable technology is a nearly impossible task in a speed-prioritizing technology environment; But that doesn’t mean some companies aren’t trying.

Immediately after the Trump administration national artificial intelligence legal framework On March 20, when “winning the AI ​​race” is paramount, technology developers face tension between shared values ​​of moving fast and breaking things, and strategically implementing responsible technology frameworks from the start.

Progress has in many cases been in the driver’s seat, and the cost of this has become clear. The difference admitted by Microsoft itself Code generated by artificial intelligence The frequent sacrifice of accessibility makes human oversight and iteration a necessity.

According to Jenny Lay-Flurrie, who took over as head of Microsoft’s Trusted Technology Group in February and has worked on accessibility for most of her 21 years at the company, responsible development and deployment of technology is twofold: “How do we make sure we’re building it right? And how do we make sure it stays right?”

Microsoft launched the Trusted Technology Group in early 2025 and has since consolidated all responsible technology initiatives under its umbrella, including Lay-Flurrie’s legacy accessibility directive.

While Microsoft centralizes its responsible technology under a top-down model, competitors like Google maintain a more engineering-focused architecture driven by core AI principles and dedicated security councils. Techniques vary across big tech, but Microsoft’s approach is one that has been reshaped since the new approach Bill Gates published in 2002. Trusted Computing grade that prioritizes things like reliability over new feature development.

Artificial intelligence’s problems (and the people who fix them)

Lay-Flurrie’s foray into the broader responsible tech space may be new, but he says it follows the same general principles he’s used to, including fairness, transparency, inclusivity and accountability. Microsoft works on this principle “humans must be responsible for AI” regardless of the consequences.

So when Microsoft realized that the AI ​​wasn’t accurately representing blind people, its team took action to fix the problem.

“Some of the images produced of blind people came back with people wearing these scary blindfolds,” he said. “These models were trained on a lot of material that exists in society. Unfortunately, society is not always the most inclusive place, so there are cases where we need to add data to train it.”

To do this, Microsoft purchased more than 20 million minutes of multi-modal data from Be My Eyes, a nonprofit accessibility platform that blind and low-vision people can use for free to connect with live volunteers and AI to provide them with audio information about what they see. “There was a lot of video material of blind people using canes and dogs and finding house keys, and we anonymized the data by blurring faces so we could more appropriately train our models on blindness,” Lay-Flurrie said.

The process is robust, but there’s room for improvement, said Annie Brown, CEO and founder of Reliabl, a machine learning training software that works to minimize bias and maximize performance in AI models.

“More diverse data is just one part of it,” Brown said. “If you don’t pay attention to what’s going on at the metadata layer—how the images uploaded to your dataset are tagged—that will create bias, too.”

Despite the world-changing AI race, Microsoft is part of a broader movement of companies publicly sharing their responsible technology learnings. Microsoft Learn is available free to students, faculty, and developers and includes training modules on the following topics: responsible artificial intelligence principles and more. Brown recommends that Microsoft also look to smaller social good organizations “to see how they are bringing inclusivity to AI.”

Lay-Flurrie says the issue of improvement comes with the territory. “Listening to feedback clearly, taking it, iterating, testing, and resolving it as quickly as possible,” he said.

People and Artificial Intelligence

Microsoft is one of the leading providers of enterprise technology; This means that its own AI fuels other companies, which often decide to reduce headcount in lieu of advanced solutions. Microsoft itself is part of a wave of major layoffs in tech, although this has been explained as a realignment of priorities rather than a complete change. company laid off approximately 15,000 people In 2025, he moved into sales, gaming and customer-facing roles, and hired new staff elsewhere focused on AI infrastructure.

Even as layoffs continue across industries, Lay-Flurrie says AI is already leveling the playing field for previously marginalized workers, including those who are neurodiverse and disabled.

“Our group of employees with disabilities was the first community at Microsoft to gain access to Copilot,” he said. “For the deaf community, captioning, transcripts, meeting notes, sign language recognition provide independence. You don’t have to wait for a cartographer to be there to transcribe what is being said.”

For the neurodiverse community that got Copilot early on, it helped with cognitive load so much that “they wouldn’t let me take the license back,” he said.

Diego Mariscal, CEO and founder of 2Gether-International (2GI), a global startup accelerator run by and for disabled entrepreneurs, agrees that Microsoft has made a breakthrough in including people with disabilities. “The fact that Jenny’s position even exists at this level is a testament to that,” he said. Still, including people with disabilities at the decision-making table is important both top-down and bottom-up. “How can we ensure that as AI develops, people with disabilities are included at the table, not from a philanthropic perspective, but because it will enable technology and innovation to become more advanced and accessible for everyone?”