Human Bias Is Everywhere in Tech. To Fix It We Need to Reshape Computer Science Education.
As we transition into the new year reckoning with a violent insurrection organized on social media, the spread of disinformation about a deadly pandemic and breakdowns in distance learning, it would be remiss of us not to acknowledge the impact of technology on every facet of our society. We all have been equally unified in our frustrations and concerns that we are inching closer to a dystopian future.
2020 shed more visibility on how racism, sexism, and other -isms permeate technology and will continue to create divides that may become irreparable. In the tech world, there have been DEI efforts, legislation targeting the racist impact of technology, warnings from ethicists and independent bias ratings created to rein in the harm—but none of these solutions address the real issue: humans. We must confront how the destructive harm unfolding through our technologies today is a reflection of the implicit bias and prejudice of the designers building the technology—and how they were taught. Many designers don’t know how to identify harmful bias and are completely unaware of how their own biases shape the products they build. So how can we start addressing this issue head-on in 2021?
Humans are the problem, and luckily education offers a solution.
We need to move beyond the quick fixes that are not working; and invest in the next generation of technologists by radically re-shaping why and how we teach computer science. The natural place to start is within the broad but influential community of computer science education, which includes teachers, administrators, curriculum designers and anyone involved in shaping how future technologists learn. Our young people need to be technically proficient in Python, R and Lisp to build AI, machine learning and other emerging technologies. However computing skills are not enough; we need to equip our young people with knowledge, skills and moral courage to design equitable tech that dismantles existing power dynamics, protects non-dominant groups, represents everyone and prioritizes the well-being of society.
As CS and technology educators we have helped create dozens of spaces for young people to tinker with technology for over a decade. Reflecting back on that time span, we can’t help but wonder how many young people graduated from those spaces capable of building a new bot, but incapable of recognizing their own biases. Where are they now? What cool and potentially dangerous technology have they put into the world? We cannot go back in time; but we can use this new insight to design a better, more equitable vision for computer science education.
A radically, reshaped computer science education will:
Prioritize racial literacy and history.
It’s important for all young people to believe they can be creators of technology, and it’s also reckless for us to omit that technology has historically been designed as a tool to surveil and oppress non-dominant communities. Prioritizing racial literacy means we must acknowledge how white supremacy has been ingrained into technology and collectively recognize that tech has not been neutral and that it has the power to harm. Examples might include how early punch card tabulators were used in Nazi Germany by the Third Reich to process racial censuses of Jewish German citizens; and how some of the first film stock centered white skin tones. Today we have technologies like facial recognition software which centers whiteness and can’t identify Black women, while also being designed to surveil and police Black and Brown communities.
Like emerging technologies, oppressive design practices have only evolved and manifested in new ways. K-12 administrators, educators and tech companies investing in computer science education need to support young people to examine the design, use and harmful consequences of discriminatory technologies.
Reflect and act on our own biases as creators.
It’s crucial for young people to understand how bias is embedded in code, data, policy and other facets of technology. The ability to reflect on how our positionality (shaped by identity and social status) influence the technologies we design is even more paramount for young people. At the Stanford d.school, we’ve built a design methodology that can help technology designers think through the first, second and third order implications of their creations before they release them into the world. Our budding technologists should iteratively evaluate their creations and ask themselves:
- Am I creating this based on my own lived experience and expecting others who are different from me to use it?
- Who benefits, who is being harmed or who is left out from the technology?
- Whose stories is this dataset telling? Whose stories is this dataset leaving out? What was the historical context when this dataset was produced?
- What don’t I know? Who should I ask and learn with?
- I can design this but should I? What are the implications that need to be considered?
Recognize and make space for multiple perspectives.
The field of design can be an arena for the “pluriverse”, which anthropologist Arturo Escobar defines as multiple ways of knowing, being and thinking that are rooted in specific places and communities.
Young people are curious, and can be inspired by the diverse ontologies and perspectives amongst the peoples of the world, and in natural systems. Guiding them to channel this inspiration into design practices which shift the power dynamics in technology across race, gender, ability and culture can make our technologies profoundly more equitable. Encouraging them to see what is possible by tackling hyper-local problems and designing solutions with others who have wildly different perspectives, is one place to start. Intercultural experiences which challenge them to question why their perspective should be the perspective held by the world, and making room for other beliefs they may not relate to, is another. These early experiences can enable them to work with others and build technologies that are more inclusive and contextually appropriate.
How do we move forward?
We can gain inspiration from the 1619 Project and the Zinn Education Project, which have provided us with the tools to face our multifaceted histories in the hopes of repairing and shaping our futures. These projects prioritize racial literacy, help young people reflect on bias and recognize multiple perspectives.
We can work with our social studies departments and across other disciplines to ensure our students have a historical understanding of technology. We can celebrate what our students code and build, and ask them to consider the impact their creations might have on others. And we can celebrate and actively engage with different perspectives that challenge dominant voices and narratives in every step of our design process.
If we can apply these practices to computer science education, our young people might create cool technology that serves everyone and upholds a just world.