Diversity Impacts Technology…But Could It Also Be Vice Versa?

My daughter inspires me to press for progress in accelerating gender equality. Born in 2017, she’s a member of Generation Alpha (2010 – 2025). The oldest members of Generation Alpha have just started school, and this generation will be the most formally educated, the most technology supplied, and wealthiest generation ever. The most striking distinction of the Alpha Generation will be their relationship with technology. They will live a life fully enhanced by analytics-driven technology, with almost every daily task made easier or more streamlined by the personalized services around them. They will also be more psychologically impacted by these services than previous generations. Sounds intriguing, right? But, what if those technologies are unintentionally biased?

Now is a critical time to talk about diversity in technology, because the people who build the algorithms and systems of today will have a significant impact on Generation Alpha and the world around us. As my pledge to #PressforProgress, I will mentor more women and girls in technology.

For my daughter and every other member of Generation Alpha, every aspect of their professional and personal experiences will be touched by analytics-driven technology services. It’s already around us in the form of search engines, security surveillance, voice-recognition personal assistants, image recognition on social networks, smart home thermostats, video captioning, music and movie recommendation products, and autonomous vehicles. Analytics-driven technology interactions and feedback are highly reflective of the algorithms, user experience choices, training data used, the creators of the algorithms, and the leaders in the space.

If we don’t have more women and minorities at the table, as programmers, engineers, data scientists and leaders, we can potentially bias systems. Attempting to reverse that ten or twenty years from now will be much more difficult, and the impact on the people who rely on them will be even more impactful. Reducing bias in analytics-driven systems comes down to having a diverse team building the products and services. Yet there’s currently an underrepresentation of women in the workforce and leadership roles in analytics and technology in general. How big is the gender gap?

  • Women make up roughly half of the workforce, but only 25.5% of computer science and mathematical professionals are women in the United States, according to the US Bureau of Labor Statistics (2016)
  • 18% of computer science grads today are women, down from a peak of 37% in the 1980’s (The American Association of University Women, 2013)
  • 40% of Statistics graduates are women (Washington Post, 2014)
  • 25% of Chief Data Officer (CDO) positions are held by women, according to a study by Triple Pundit (2016)

Why does the tech gender gap exist? Unfortunately, women still grow up hearing directly or inferring through signals around them that “technology professions aren’t for girls”. They don’t get as much encouragement, they lack strong role models, and they even receive negative peer pressure.

Without increased diversity, we risk locking biases into technology that may impact future generations. To build the personalized services consumers expect, analytics teams may train analytic models to identify images, recognize voices, filter content, or spot safety hazards using large data sets. The model “learns” how to apply the machine learning algorithm from a training data set: of photos, videos, or a database of numbers that lays the groundwork for its functionality. But if that training set is biased in some way, that’s what the AI learns is normal – ultimately, what it produces reflects the data provided.

Take image recognition, which may use machine learning to identify images of a leader. First, the data scientists must identify pictures of leaders. Then they identify numerous pictures that are not leaders. They pass a deep learning algorithm over the images to learn to accurately predict whether the image is a leader. When provided with a new image, the model will answer with the probability that the image is a leader. Now, imagine if the people selecting the data sets and training the algorithm were not diverse, and/or were not trained to spot unconscious bias in algorithms.

Also important is that many of the focus areas for analytics research and investment will be impacted by a lack of diversity. If the leaders and the teams choosing the problems to focus on are not diverse, they are less likely to prioritize the types of problems faced by the majority of the world. When problems don’t impact us, we are less likely to focus on solving them, right?

We need more women working in technology, preventing bias in analytics algorithms, and helping to develop more innovative solutions. We need more female data scientists and female programmers, and we also need more women elevated into leadership positions in technology. There’s plenty of research that demonstrates that having more women at almost any level of your company, and especially in leadership, has a positive impact on results and a company’s bottom line. Further, research has shown that more diverse leadership teams develop more new products, and turn more of that innovation into revenue.  

What can you do to help close the gender gap in technology?

  • CELEBRATE the accomplishments of women in technology
  • RAISE your voice against discrimination and unconscious bias
  • MENTOR and sponsor more women and girls
  • SUPPORT women’s networks, to increase support and inclusivity for women in tech
  • EXPAND the pipeline by encouraging and educating the next generation of women in STEM

Just as we are working to prevent bias in hiring and promoting people, we need to focus on getting more women into technology to ensure we limit bias shaping the technology that will build our collective future. We also need more women in tech today to inspire and motivate the girls of Generation Alpha to follow our lead for a brighter tech industry of tomorrow.

Morgan Vawter

Chief Analytics Director