There is a good chance you are reading this article on a laptop, tablet, or smartphone. There is an even greater chance that thanks to a woman, your piece of electrical hardware is able to function at such a seamless rate you would hardly even notice a glitch. However, there is a major glitch behind-the-scenes of the appliance at your fingertips: the systemic oppression of female coders.

“I’ve been called a ‘baby’ because I made a mistake...I just don’t see a dude CEO calling another dude a baby,” said Maddie Shang, a research engineer specializing in interactive and reasoning machine learning that can track data on healthcare communities or economic marketplaces. “There is still so much malicious bias because you are a woman. Some male coders just hate you because of their unconscious bias.”

Is Shang the only woman who has dealt with bigotry based on her gender in the world of artificial intelligence (AI)?


A few years ago, a similar experience halfway across the globe in India slammed into the consciousness of undergraduate student Kritika Prakash during one of her math courses at the International Institute of Information Technology (IIIT) Hyderabad. IIIT Hyderabad is considered one of the premier computer science universities in the country. As one of many exemplary engineering students, Prakash was lucky to have even a few female colleagues in her lectures. Such was the case in one particularly challenging math class of 125 students on graph theory.

Until things changed.

After the first round of exams, 100 students dropped the course. Leaving Prakash as the only woman in the class. “It was extremely challenging,” Prakash remembered. “Almost all my friends had left it by then and I was surrounded by a bunch of guys who had these ideas that they were really smart and so they didn’t need to work hard to get results.”

Prakash’s peers actively discriminated against her because she was a woman. Her XX chromosomes led her colleagues to voice their opinions about how at best she would get an average grade, but more likely would struggle through the material.

Prakash proved them all wrong.

“At the end of the semester, I had the highest grade in the class...that was great.” Prakash said. “That was a huge confidence boost and a lot of fun.”

Her favorite memory involved the advanced graph theory final exam which only had three questions. Almost everybody in the class could solve one of them. “I was one of the only ones who could solve two,” Prakash remembered with a laugh. She believed her success was not a result of her intelligence, rather her increased level of effort in comparison to her male classmates.

So what is the problem both Shang and Prakash have had to face? To state it plainly, women struggling with internalized imposter syndrome in a patriarchal community. Gender oppression is not unique to the world of AI. Just check the record on the past thousand years of history. However, the technology landscape is changing - rapidly. When it comes to the exchange of online information, the ability for companies and coders to adapt is crucial and complex. Yet, one group in the world of data privacy is facing the problem head on: OpenMined.

OpenMined is a non-profit organization with a mission to create an accessible ecosystem of tools for private, secure, multi-owned governed AI. In layperson terms, they are trying to make it easier for coders to have access to large amounts of data which will improve their computer algorithms. This would allow for better efficiency in hospitals, classrooms, and other tech projects which affect our everyday lives. Far too often, only multi-million dollar corporations or institutions have the means to create algorithms that can actually make a difference in the world. Yet, OpenMined has the vision that any person with a laptop should be able to create constructive code. Particularly in a safe manner that protects people’s private information.

“Privacy is actually about the norms of information flow,” said Andrew Trask, leader of OpenMined. “It’s about where information is flowing and in what context. That is news to the world a little bit because we have social norms and social expectations about who should know what and in what context. Privacy technologies is about making that flow happen.”

This open-source community of coders and machine learners all sharing their best information has been deemed a Digital Public Good by the United Nations. The goal of lowering the barrier-to-entry to private AI technologies is part of the reason Shang and Prakash have joined OpenMined’s initiative. Shang is now the OpenMined Team Lead of the Recommender Systems Team. While Prakash is the OpenMined Team Lead of the Differential Privacy Research Team.

“OpenMined is just naturally a very friendly and encouraging community,” Prakash said. “It has honestly helped me a lot in being more confident and getting over that feeling of imposter syndrome.” Both Shang and Prakash will be sharing their personal experiences within the coding community and their work at OpenMined’s global conference on privacy, adaptly named #PriCon2020.

The conference is this weekend from September 26-September 27 and is hosting more acclaimed female coders than just Shang and Prakash. The head of Microsoft’s Cryptography and Privacy Research group is Dr. Kristin Lauter who will be presenting. As will Dr. Helen Nissenbaum who is credited with creating the notion of data social norms and the frameworks of privacy. Another female superstar in the field is Cynthia Dwork who co-invented the Differential Privacy sector.

“Quite naturally, the biggest names in privacy are all women which is quite wonderful,” Trask said. “Maybe it is just the nature of privacy itself, maybe it attracts people more broadly, maybe there is a better gender distribution because it is focused on social good as opposed to strategic gains.” Trask highlighted how the philanthropic socialization of data privacy might be a key player in making it a more female friendly arena. “There are a lot of natural female role models that attract more female participation which is fantastic.”

While mentors are pivotal in changing the perception of an industry, Prakash can’t help but eventually hope for a world where her identity as a female coder is obsolete. “The goal is to not have someone ask me what it is like being a female in anything. Eventually, that should never be brought up,” Prakash said. “Hopefully I don’t need to speak on my own behalf, my work can speak for me. That has been my approach for a long time. Ignore what people are saying, just work so hard that people can tell things by your work.”

Unfortunately, for some female coders, a strong sense of drive is sometimes not enough. “The sad truth of the matter is that there are a lot of very undervalued women who want nothing but to do their best work and contribute as much as possible to whatever they are working on, except, for whatever reason, they don’t have the opportunity to do that,” Shang said. A year ago, Shang founded her own professional group called “Women Who Code Data Science,” which boasts over 2,000 members to try and increase favorable circumstances for women.

Despite hundreds of thousands of initiatives across the globe like Shang’s hoping to encourage female participation in coding, America has hit historically low percentages in the past few years. In 2016, only 18.7% of undergraduate women declared a major in computer science as opposed to 28.4% in 1994, according to the American Enterprise Institute. Considering the increased awareness surrounding the need for gender diversity in STEM fields over the past three decades, assumedly that percentage would have increased.

“The tech industry is constantly saying there is a ‘talent gap’ or a ‘diversity gap.’ Well, you’re looking in all the wrong places,” Shang highlighted. “Whatever you’re doing for your recruiting or trying to get more diversity in STEM is not working. Let’s look at what works, try and fix some of your problems, and maybe women who are hardworking and intelligent will get their shot.”

Bringing more women to the table has not been so defeating across the STEM board. In 2019, for the first time in American history there were more women in medical school than men, according to the Association of American Medical Colleges. So, why is that not the case when it comes to computer science?

One possible answer is increasing visible representation and breaking down financial barriers. A solution, OpenMined and #PriCon2020 have both deployed. “If you’re looking at the pipeline for machine learning researchers in places like Silicon Valley, they are pretty much always looking for someone with a PhD, someone who has published papers, or someone who is already a part of an industry team,” Shang said. “What are the people at conferences and paper reviewers looking for? The fact that you have a PhD. The fact that you don’t have an advanced degree becomes the bottleneck.”

While having a PhD in computer science or mathematics may be the norm for many industry tycoons, #PriCon2020 never made an advanced degree a prerequisite to present your work. It’s not even an unspoken rule of thumb. Which is why Shang, Prakash, and so many other speakers have been invited. They are coding with a purpose regardless of whether or not they are earning a doctorate.

“This sounds like the numbers don’t matter but they really do and here is why,” Shang said. “As people, we are all more attuned to the problems that affect our communities more. It’s not that the researchers don’t care, they try really really hard to make sure they are doing good work and hopefully not doing damage. But if you come from a certain background how are you supposed to understand how this will affect a different community or someone from a different background?” Too often, computer algorithms include the implicit bias of society which marginalize minority communities. “Therefore the only way, and the most important way, we can make sure the machine learning we build today will help greater humanity rather than just the few and the privileged is to have representation at every level.”

Money should also not be a barrier when it comes to improved machine learning. A fact OpenMined has taken to heart with a number of free instructional coding courses throughout the year. Plus, getting the chance to see so many female and male role models alike at #PriCon2020 is 100% free. For more information on the #PriCon2020 Agenda and Free Tickets visit:

“If we are trying to encourage this idea of fair and diverse representation in machine learning research so that we can try to make algorithms that serve greater humanity for each community, this bottleneck is something that we really need to think about,” Shang emphasized.

Trask has heeded cries for technological justice like Shang’s and with his team has recently adapted the goal of OpenMined. It is an objective that involves more than just privacy, rather opting for a more human-centered coding community. Trask will reveal the new mission statement during his talk “High-Level Technology Roadmap” at #PriCon2020.

Getting to hear from some of the leaders at OpenMined and in privacy like Lautner, Shang, or Prakash will draw you in regardless of your familiarity with computer programming. It will be in support of the future which is looking more and more female.

“Even if you’re not interested in machine learning and not a very technical person, machine learning is already in your life and it will take up more and more of your life as automation increases,” Shang emphasized. “Your data is currently being used. Even if you don’t care about it, it does affect you. This is something I hope people will get vocal about, get mad about, to help inform researchers, academia and companies to try and drive change.”

So while the problems of oppression might seem challenging and difficult to change, OpenMined seems to have found at least one possible solution: elevate the profile of female role models and the support of all women within the machine learning community. Then, let some of the best problem solvers in the world do the rest.