Sowmya Subramanian ’96 examines how tech can empower women

Photo by Margaret Skelly ‘21  Sowmya Sumramanian as she discusses Freedom of Expression

Photo by Margaret Skelly ‘21

Sowmya Sumramanian as she discusses Freedom of Expression


“The internet has enabled freedom of opportunity,” said Sowmya Subramanian ’96, senior director of engineering at Google at her March 6 talk in Gamble Auditorium, though she added that there is often a gender gap in terms of access to information and resources.

Subramanian presented to Mount Holyoke community members on Wednesday evening, speaking about “Empowering Women Through Technology.” According to the College website, “In her dozen years at Google, Subramanian has founded and led YouTube Kids, YouTube Music and YouTube Live. She has been critical in expanding YouTube’s revenue stream from an ad-based model to a paid-subscription model, known as YouTube Red, and transformed YouTube into one of the top global platforms for music artists and fans.”

In addition, “prior to YouTube, Subramanian led engineering in Google Maps, where she defined and executed on its mapping data and local business efforts, and won several Google leadership awards. Subramanian co-leads Women@ for Google North America” and “serves on the advisory boards of PBS Kid, iTVS Women and Girls Lead Global.”

The event was sponsored by the McCulloch Center for Global Initiatives and the Department of Computer Science. Professor of Economics and Carol Hoffmann Collins Director of the McCulloch Center for Global Initiatives Eva Paus introduced Subramanian, standing next to a big, blue Mount Holyoke banner hung over the auditorium’s chalkboard.

Subramanian began by speaking about her experience with YouTube. “YouTube started with a very simple mission,” she said. “It was just simple. Broadcast yourself.” She said that thanks to a huge transformation in technology since the website first began, YouTube now sees 400 hours of videos uploaded each minute.

She spoke about the relationship between women and the internet, citing examples of women who have utilized the internet to start businesses, escape abusive relationships, advocate for women’s education and seek asylum in countries around the world. Even though women have been able to use the internet to their advantage in these cases, she explained that there is still a gender gap in how technology is being made available to men and women.

Subramanian highlighted Project Loon, an effort to expand global internet access using stratospheric balloons created to bring internet connectivity to hard-to-reach places around the world. The goal of this project is to expand internet access for everyone.

Subramanian said that there is also a lot of intersection between healthcare and technology, another area in which it is important to consider gender. For example, many clinical trials and medical tests are skewed male, according to Subramanian. This can potentially lead to a neglect or lack of understanding of treatment options that work well for women but may not for men.

Fairness in machine learning, which enables technology like voice recognition, navigation and language translation, is important for reducing this kind of bias, according to Subramanian. This type of artificial intelligence is made possible by computers finding patterns in data, but Subramanian said, “just because something is based on data does not make it neutral.”

She showed a video by Google titled “Machine Learning Fairness” to illustrate her point. The video told the audience to “close your eyes and picture a shoe.” The screen then displayed the image of a sneaker, a traditionally men’s dress shoe and a high heeled shoe. The image selected would depend on what a person teaching the computer to recognize a shoe already had in mind. “This is how bias happens in machine learning,” the video said. “This is how bias happens in machine learning.” This can be a problem when gender, race, culture and other identities are not considered when developing technology. According to the Google developer webpage on Machine Learning Fairness, “Building inclusive machine learning algorithms is crucial to help make the world’s information universally useful and accessible.”

Subramanian went on to discuss her work with YouTube and Google on improving online safety for women. She referenced the disproportionate harassment of women gamers through hashtags and disparaging comments online, as well as inappropriate comments on YouTube videos of women and children. She explained that YouTube actively works to develop tools which moderate comments and aim to reduce predatory online behavior. “With online safety, every day we are learning something new,” she said.

In an effort to reduce stereotyping online, tech companies like Google have also worked to diversify emoji options, giving them customizable hair and skin colors to promote inclusivity.

Expanding options is something Google and other companies have done, as many people may notice when choosing emojis to add to their text messages or emails.

She also discussed the importance of improving gender considerations in music, media, hiring and education. Following the talk, students asked questions about a variety of topics, from increasing access to information for women to addressing bias online.

Sarah Caggiano ’19, a computer science major and math minor, attended the talk. “I definitely liked that she was an alum,” she said. “It’s nice to see someone from Mount Holyoke succeeding in such a short amount of time in such a big company.”