Women in AI speaker series offers insights into the importance of diversity in tech
In collaboration with Deloitte, the Schwartz Reisman Institute for Technology and Society presented a distinguished speaker series highlighting contributions from six innovative female researchers working at the forefront of artificial intelligence across a wide range of specializations. Left to right: Anna Goldenberg, Joanna Batstone, Cynthia Rudin, Arisa Ema, Sophia Ananiadou, Maren Bennewitz.
Designing ethical and trustworthy artificial intelligence (AI) systems that benefit all of society requires a diversity of perspectives working together in equitable alignment.
To promote equality and inclusion and champion the success of female-identified researchers across the field of AI development, the Schwartz Reisman Institute for Technology and Society (SRI) partnered with Deloitte to present “Women in AI,” a speaker and mentorship series designed to connect attendees from around the world with distinguished female thought leaders in the field of AI research.
Over 1000 attendees from around the world joined virtual sessions on topics ranging from creating AI-powered prediction and decision-making tools for healthcare to building an effective AI governance ecosystem.
The event series was part of Deloitte’s Women in AI – Global AI Leadership & Development program, which ran from October 2022 to April 2023, and is part of a suite of diversity, equity, and inclusion programs run by Deloitte. This first-of-its-kind AI leadership program was designed specifically to bring Deloitte AI leaders, leading academic researchers, industry executives, and Deloitte women working in AI together to share their experience and advice– as well as their AI research and models.
Over a six-month period, the Women in AI series featured speakers who discussed their diverse research and experiences working in a largely male-dominated field, highlighting the innovations they have brought forward in their respective areas of expertise and showing how their work is helping to change the culture of AI research. The series also featured closed mentorship sessions for members of the Deloitte Women in AI Global AI Leadership Development Program.
Kelly Lyons (left), professor in the Faculty of Information at the University of Toronto and a faculty affiliate at the Schwartz Reisman Institute, co-hosted the Women in AI series with Aisha Greene (right), manager of the Canada Deloitte AI Institute.
Uniting to create change for women in technology
“Change and innovation does not happen in isolation,” says Beena Ammanath, partner at Deloitte. “For this reason, we are engaging with the broader community, from non-profit organizations to leading female AI experts and change makers to create spaces where the next generation of AI leaders can thrive and grow.”
“This type of initiative is truly a communal endeavour,” says Monique Crichlow, executive director of the Schwartz Reisman institute. “We rely on not only the expertise and generosity of the invited speakers and the hard work of organizers from both Deloitte and SRI, but also on the curiosity, commitment, and ingenuity of the attendees. They asked questions, they networked, and they created lifelong connections that will only strengthen the community of women supporting each other—in tech and beyond.”
What advice do women in AI have to offer?
From Anna Goldenberg, a University of Toronto computer scientist working with clinicians to build AI-powered prediction and decision-making tools for healthcare to Joanna Batstone, a physicist-turned-data scientist who is the inaugural director of the Monash Data Futures Institute, the series’ speakers represented a wide variety of experience, both professional and personal.
Across all six talks, one recurring piece of advice continued to ring: have confidence.
“So many women take themselves out of the equation, out of the competition,” said Goldenberg. “It’s OK not to be the smartest in the room and still speak up. Always know that your perspective is an important one to contribute.”
Batstone echoed Goldenberg’s advice with a practical tip on navigating professional interactions with confidence. “I encourage you to have a 30-second elevator pitch that acts as a summary of what you’re working on and why it’s important,” she said to attendees. “In whatever situation you’re in, you should be able to clearly and succinctly describe and defend your work. Everyone will have an opinion, so you should have your bullet points ready to state what AI means to you in the context of the projects you’re working on.”
New paradigms for discovery and discussion
In her talk, Duke computer scientist Cynthia Rudin presented her research on simple machine learning models, questioning whether complex, proprietary models are in fact superior to simple, interpretable ones—or whether these “black box” models are even necessary at all. The implications of Rudin’s findings are potentially transformative for high-stakes decision-making contexts, like criminal justice and loan decisions, where issues around the explainability of AI systems are of great importance to both those who deploy and those who are impacted by them.
In addition to reducing bias in predictive models, Rudin proposed a new paradigm for the world of machine learning in which we hand users (experts in their domains such as doctors or law enforcement) a number of different models, “so they can pick something that doesn’t just agree with the data, but also agrees with their domain knowledge.”
Rudin relayed how she received a lot of pushback early in her research career.
“At one point, the whole idea was that machine learning should do everything for you and you shouldn’t have to have a human looking at anything,” she observed. “I disagreed with that, and it was controversial. Many people didn’t like it, but now they’ve started to come around. It was really hard to get papers into any publication venue.”
Rudin was recently awarded the distinguished AAAI Squirrel AI Award for pioneering socially responsible AI for her groundbreaking work in interpretable AI systems, and she is a three-time recipient of the INFORMS Innovative Applications in Analytics Award.
Responsible AI expert Arisa Ema from the University of Tokyo gave attendees advice on how to acquire buy-in and encourage collaboration when building an AI governance ecosystem.
“It can be difficult to involve industry people in discussions [about responsible AI] because ethical or legal issues in AI are not raised as much as innovation,” said Ema. “However, I think things have been changing.”
Her advice?
“Start from cases and be concrete. Demonstrate the risk scenario. Show how we can identify risk and collaborate with each other. While it’s important to talk about principles like fairness and transparency, what we are doing instead is creating scenarios that demonstrate how everyone’s cooperation is needed in solving these problems.
Batstone’s talk also touched on differences between conducting research in industry and the academy, and she encouraged women not to shy away from asking important questions when navigating their careers.
“For any of us considering a career move, look at the organization you’re considering joining in the context of the quality of the people in the organization. Is it a team of people that are leading, respectful, kind, and have a focus on looking at ways we can do things differently in the future?”
The series was rounded out by talks from Maren Bennewitz, a professor for humanoid robots and vice rector for digitalization at the University of Bonn and Sophia Ananiadou, professor of computer science at the University of Manchester and director of the UK National Centre for Text Mining. Bennewitz presented her latest research on robots acting in human environments, while Ananiadou spoke about her work using text mining and natural language processing to facilitate the discovery of new knowledge, focusing on how semantic search of biomedical literature can quickly review health and safety incident reports and identify risk in the construction industry. Ananiadou noted that her lab is staffed by more than 50% women, demonstrating how female leadership can generate organizational shifts away from inequities in the field.
Participants in the Women in AI program reported an enriching and empowering experience. “This program dedicated to developing women leaders in AI shows the commitment that Deloitte has to inclusivity,” says Riona Arjoon, Senior Manager in Consulting at Deloitte in the Netherlands. “It was great to focus on content as AI leaders as well as soft skills and best practices from inspirational women globally.”
Want to learn more?
Watch video recordings from all six events in the Women in AI series.
Read about the recently-launched Deloitte Women in AI Global Leadership Development Program.
Learn more about Deloitte AI Institute’s ongoing research projects.
Explore all upcoming events hosted by the Schwartz Reisman Institute for Technology and Society.
