As a part of TechCrunch's ongoing Women in AI series, which goals to offer AI-focused female academics and others their well-deserved (and overdue) time within the highlight, TechCrunch interviewed Sophia Velastegui. Velastegui is a member of the National Science Foundation (NSF) National AI Advisory Board and former Chief AI Officer of Microsoft's enterprise software division.
Velastegui didn’t plan to pursue a profession in AI. She studied mechanical engineering as an undergraduate at Georgia Tech. But after a job at Apple in 2009, she became fascinated with apps — especially AI-powered ones.
“I began to see that AI-powered products were resonating with customers due to the sense of personalization,” Velastegui told TechCrunch. “The possibilities for developing AI that might improve our lives on small and huge scales seemed countless, and I desired to be a part of this revolution. So I began searching for AI-focused projects and took every opportunity to expand from there.”
AI Forward Career
Velastegui worked on the primary MacBook Air – and the primary iPad – and was soon named product manager for all of Apple's laptops and accessories. A couple of years later, Velastegui moved to Apple's special projects group, where she helped develop CarPlay, iCloud, Apple Maps, and Apple's data pipeline and AI systems.
In 2015, Velastegui joined Google as head of silicon architecture and head of the Nest brand product line. After a temporary stint at audio technology company Doppler Labs, she accepted a job offer at Microsoft as general manager of AI products and search.
At Microsoft, where Velastegui eventually led all AI initiatives related to business applications, Velastegui led teams in adding AI to products resembling LinkedIn, Bing, PowerPoint, Outlook and Azure. She also led internal explorations and projects built using GPT-3, OpenAI's text generation model for which Microsoft recently acquired exclusive license.
“My time at Microsoft is actually special,” said Velastegui. “I joined the corporate when it was within the midst of major changes under the leadership of CEO Satya Nadella. Mentors and colleagues discouraged me from taking this leap in 2017 because they believed Microsoft was lagging behind within the industry. But inside a brief time frame, Microsoft had began to make real progress in AI and I desired to get entangled.”
Velastegui left Microsoft in 2022 to start out a consulting firm and lead product development at Aptiv, the automotive technology company. In 2023, she joined the NSF's AI Committee, which works with industry, academia and government to support basic AI research.
Navigate the industry
When asked how she overcomes the challenges of the male-dominated technology industry, Velastegui cited the ladies she considers to be her strongest mentors. It's necessary for ladies to support one another, says Velastegui – and, maybe even more necessary, for men to get up for his or her colleagues.
“If women in tech have ever been involved in transformation, adoption or change management, they’ve the fitting to have a seat on the table. So don’t be afraid to take your home there,” Velastegui said. “Raise your hand to tackle more AI tasks, be it a part of your current job or a large-scale project. The best managers support you and encourage you to maintain moving forward. But if that’s impossible in your on a regular basis life, search for communities or university programs where you may be a part of the AI ​​team.”
An absence of diverse viewpoints within the workplace (e.g. AI teams which can be predominantly male) can result in groupthink, notes Velastegui. That's why she advocates that ladies give feedback as often as possible.
“I strongly encourage more women to get entangled in AI in order that our voices, experiences and viewpoints are included at this critical start line that may define foundational AI technologies for now and the long run,” she said. “It is crucial that ladies in every industry are genuinely all for AI. If we join the conversation, we will help shape the industry and alter this power imbalance.”
Velastegui says her work now at NSF is targeted on addressing open fundamental problems in AI, resembling the dearth of what she calls “digital representation.” She claims that today's AI is riddled with bias and bias, partly as a consequence of the homogeneous composition of the businesses that develop it.
“AI is trained using data from developers, but developers are largely men with specific perspectives and represent a really small subset of the world’s 8 billion people,” she said. “If we don’t include women as developers and if women as users don’t give feedback, then AI won’t represent them in any respect.”
Balance between innovation and security
Velastegui sees the rapid pace of the AI ​​industry as a “huge problem” – that’s, the dearth of a standard ethical security framework. She believes that such a framework, if ever widely adopted, could enable developers to develop systems quickly without stifling innovation.
But she doesn't expect it.
“We have never seen such a transformative technology evolve so rapidly,” Velastegui said. “People, regulations, legacy systems… nothing has ever had to maintain up with the present speed of AI.” The challenge is to remain informed, current and forward-thinking while being aware of the risks of moving too quickly.”
How can an organization – or developer – develop AI products responsibly today? Velastegui takes a “people-centered” approach, learning from past mistakes and specializing in user well-being.
“Companies should empower a various, cross-functional AI council that reviews issues and makes recommendations that reflect the present environment,” Velastegui said, “and create channels for normal feedback and oversight that adapt because the AI ​​system evolves. “And there must be channels for normal feedback and oversight that adapt as AI systems evolve.”
