Last year, 2018, was a milestone year for women. Industries ranging from tech to entertainment, and everything in between, took a hard look at issues of inclusion regarding women, in a movement championed by brave members of that gender who pointed out the many areas in which we still lag behind.
The results were powerful. Over the last year, we’ve seen legislative moves, such as California’s new law requiring women on corporate boards; initiatives like Melinda Gates’s Reboot Representation Tech Coalition; and diversity talks from compelling advocates like Kesha Williams, the software engineering manager at Chick-fil-A, about the goal of boosting women’s representation.
In my field, artificial intelligence (AI), we witnessed our industry’s most prominent conference grapple with issues of insensitivity and jokes made about sexual assault, before eventually changing the event’s very name itself because of its offensive connotation for women.
Headlines like these gave me hope. But as a female executive in a male-dominated industry, I can’t help but continue to question what comes next. Women in STEM are grossly underrepresented — making up just 26 percent of the computing workforce for instance — and make on average $16,000 a year less than our male counterparts.
So, setting parameters or quotas was a good starting point for inclusion, but we need to take things a step further. How can women, and our male allies, continue to chip away at inequality?
We need to re-examine how the tech industry — and workplaces in general — address differences, and ensure diversity is intrinsic in the technology we build. Here’s how.
Don’t eliminate “gender” from the conversation. Don’t ignore “identity.”
Before we can address representation in the companies we lead and work for, we need to look at how the tech industry handles differences at its highest level. Faced with potential gender issues, many tech companies are quick to try to eliminate gender from the conversation entirely. For example, Google recently removed gendered pronouns from Gmail’s Smart Compose feature to avoid AI bias. The sentiment was genuine, but the approach missed the mark.
In fact, we can’t ignore people’s identities — including gender, race, ethnicity, age, background and the like — because these things are core to our humanity. We need to find a way to recognize and celebrate all of these differences, in the technology and products we build, and in the companies we lead, rather than erasing them for the sake of simplicity. If we encourage everyone to be the same, then the technology we build will only serve a small population and fail to serve the majority of people in the world.
This is especially important in the industry that I am in. Artificial intelligence has the potential to significantly improve lives, but it’s also in danger of widening the bounds of inequality across gender, race and other characteristics. This is an issue transcending the tech space and reaching other corners of society already. Just recently, U.S. Rep. Alexandria Ocasio-Cortez, D-N.Y., addressed the issue of algorithmic bias during an event on MLK Day.
If AI is developed by only one specific group of people, it will serve only that group. One way to prevent this outcome is to ensure that the data that AI is built on represents different kinds of people. It can be challenging to amass diverse data, but we’re taking steps as an industry.
For example, IBM recently released a dataset, “Diversity in Faces,” which aimed to reduce biases in AI facial-recognition tools. The dataset provides the public with an annotated, diverse dataset of one million human faces — the largest dataset of its kind — making diversity readily accessible for developers to build into AI.
Find, or create, the right tools.
Another way to prevent technology from perpetuating inequality is to ensure that people have equal access to the tools they need to develop tech that works for their communities. My team has developed our Brain Builder platform with that in mind: We’re striving to give developers, students and universities, regardless of gender, race or identity, more cost-effective and efficient tools to develop AI.
We’re also able to give more insight into the data behind AI systems, to produce explainable AI: solutions that can be trusted, easily understood and therefore more easily judged, to ensure fairness and accountability. In turn, we hope to mitigate biases.
In another example, Amazon created an experimental AI tool to expedite the hiring process by giving job candidates scores ranging from one to five. Soon after, the company realized that software development and other technical positions were being scored unfairly, and were subject to gender bias. Amazon’s computer models had been trained to vet applicants by analyzing resumes submitted to the company from the previous 10 years, and most of those resumes had come from men!
This example drives the point home: Aside from fighting for representation in the workplace, we need to consider representation in the technology we build, too. That way, the systems we put out into the world will serve as another tool for ending inequality.
Hire consciously, and champion male allies.
A lot of the conversation around representation calls for more women and people of diverse backgrounds at the table. In turn, we need to recognize our male allies and make sure they’re present, too.
Many men advocate for women in decision-making spaces — not because they need to check a box, but because they recognize the value of having diversity at the table and the expertise women offer. True (male) advocates, of which there are many, recognize that different backgrounds and identities improve an organization for everyone, including themselves.
In fact, in 2018 the Financial Times recognized a group of 50 male executives who’d made strides for women in business with initiatives like Bloomberg’s Business of Equality Summit and BNY Mellon’s EMEA Diversity and Inclusion Council. We, as women, need to celebrate and listen to our male allieswho also value workplace diversity.
Recently, I was at a dinner with female colleagues in the AI industry. One point we agreed upon was that, unfortunately, there are still spaces and conversations where female leaders are severely underrepresented. A lot of these conversations that are starkly male pertain to money — funding, venture capital and the like, which are crucial to business decisions.
By some estimates, a staggering 98 percent of VC funding goes to men, largely due to the fact that women are underrepresented in VC. We need to find ways into these conversations; but with so many of the decisions still happening among men, we aneed to ensure that we have allies in the room when we’re not present.
Cultivate a culture of challenging conversations, openness and honesty.
One thing I’ve realized in being a female leader is that you can’t shy away from tough conversations. Biases exist, and social norms and stereotypes need to be chipped away at.
For example, questions surround how to handle motherhood for women in the workplace. I’m no stranger to this, as a (soon-to-be) mother of four. I’ve found that people still assume that women won’t be able to do things, but we need to refocus on what women can do. Other conditions that require a change in work are not judged, so pregnancy shouldn’t be either. Nor should child care. Women need to feel safe and supported to discuss these things.
Support motherhood at work.
Some companies are already doing great work to make mothers feel welcomed and supported. Salesforce offers a number of parental benefits designed to support motherhood through generous parental leave, access to breast milk shipping for new mothers traveling for business and even counseling to help families with the college admissions process. Such programs are key to creating a culture of inclusion and normalizing working mothers as key contributors to an organization.
The bottom line is, we need to call out patterns of bias and stereotypes (in our companies, industries and society at large) and unpack them. It’s not about embarrassing people, but starting conversations at a micro-level, within our own companies, which have a ripple effect over time.
Check all judgments at the door.
In 2019 and beyond, ensuring women — and people of different backgrounds — a seat at the table needs to be more than just a checked box for “diversity.” Women have valuable expertise and perspectives, and they should be included, recognized and compensated based on those merits, identity aside.
So I challenge men and women alike to consciously check their judgments and perceptions at the door. The tech industry, and society at large, has a lot to gain if we allow for the influence of the many diverse voices and perspectives in our orbit.