Melissa Valentine: Understanding How Companies Best Incorporate Machine Learning
Management researcher Melissa Valentine has never been content to stay in one lane; her interests lie in trying to make sense of the competing forces and unpredictability that make up organizations of all kinds.
“I think ‘interdisciplinary’ explains a lot of my trajectory,” she says. “I see the system, how different groups are interacting. It’s the way my brain works.”
The Stanford associate professor of Management Science and Engineering is a faculty affiliate at the Stanford Institute for Human-Centered Artificial Intelligence (HAI) and its first sabbatical scholar. Her road to HAI was – for her – typically unpredictable. It began with undergraduate work at Stanford in human biology.
“Here, that includes a lot of social science along with the biological and medical science,” she says. “I started learning about quality-of-care issues in healthcare, and the more I learned, the more I realized that these were organizational issues. When I found out there was an academic discipline that focuses on organizational problems, I became really interested in organizations as systems that could be improved.”
Valentine pursued that interest for her PhD at Harvard Business School, eventually taking a leap of faith when it came to finding her next role.
“I was interviewing at business schools, but when I learned about Stanford Engineering’s Management Science and Engineering Department and its focus on studying management from a technological point of view, I interviewed here, and they hired me,” she says. “I wasn’t really focused on technology at that point, and most people with my background don’t sit in engineering schools. But once I was here, I met a group of computer scientists who were using a platform to study team coordination, which is what I’d done my dissertation on. It started a whole new research agenda for me. I’m so grateful I came here, because it’s shaped my whole career. Everything I do now is based on how technology is shaping work.”
Valentine currently has two primary areas of interest. One examines how artificial intelligence in the workplace is impacting human expertise. She recently embedded with a fashion company for 10 months as part of a project that asked experienced buyers to work with data scientists to measure how successful their purchase decisions actually were.
“These buyers began to use AI to test their intuitive ways of thinking and their purchasing decisions, and sometimes found that their own decisions were incorrect,” she says. “They became experts at using AI to test their old expertise, which actually created a new type of expertise.”
This augmentation of expertise by AI has the potential to be both empowering and threatening, Valentine says.
“AI is changing the nature of expertise, which is a profound thing,” she says. “People will have to grapple with what this new expertise means for who they are, and for what it means to be good at their jobs.”
Valentine has also teamed up with Stanford Assistant Professor of Computer Science Michael Bernstein to study flash teams, temporary organizations of employees who are convened through online platforms, work together remotely, then disperse. Massive online labor markets are making it easier for individuals and organizations to find experts, hire them quickly, and coordinate their interaction on increasingly complex projects.
“We get a lot of ideas from people about how you could use flash teams to innovate and solve problems,” Valentine says. “One compelling idea someone sent us was a platform that quickly and intelligently convened groups of doctors for remote discussion on complex diagnoses, rather than their standard process of paging and waiting for consults in more ad-hoc ways. They were thinking of how to use technology and the flash team approach to speed up expert group decisions instead of waiting to schedule meetings with everyone in the same room to talk.”
Like AI-augmented expertise, flash teams have the potential to impact employees and to change how organizations work, and as such must be carefully considered, she says. Some larger companies have begun experimenting with algorithms for in-person work, putting out calls for employees willing to come in for an hour or two of work when staffing is lean, for example.
“Some people will find the idea a positive one, while for others it may be more precarious,” Valentine says. “Would these workers always have to be searching for a job? What about healthcare and retirement benefits, and access to community? We need to start thinking about social needs and protections of those working within these systems.”
Stanford HAI is one of the places best-equipped to tackle that work, Valentine says.
“HAI intersects with so many different disciplines,” she says. “Having a chance to spend time with all these technical and interdisciplinary people helps me dive deeper into my own interdisciplinary work, and allows me to hear the questions that businesses are bringing to HAI. I hope I’m also helping others here think about what’s going on in business and management in the workplace.”
In the near future, Valentine plans to study how data science can be used to better structure the workplace, and how machine learning can most effectively be introduced into large, long-established companies.
“I’m really interested in organizational effectiveness, and AI technology has the potential to help organizations solve problems and do a better job,” she says. “But I also love to study people who are solving problems, and I respect and admire the way people create meaning in their lives and their jobs. Work is where people are, and sometimes it’s where they make the contributions they’re most proud of.”
This article is part of the People of HAI series which spotlights our community of scholars, faculty, students, and staff coming from different backgrounds and disciplines.
Stanford HAI's mission is to advance AI research, education, policy, and practice to improve the human condition. Learn more.