The 2020 documentary Coded Bias highlights an urgent problem: artificial intelligence and facial recognition technologies, often marketed as “neutral” and “objective,” in fact reproduce—and sometimes intensify—racial, gendered, and other social inequalities.
At the center of the film is MIT computer scientist Joy Buolamwini, who discovered that the facial recognition software she was testing could not detect her face unless she wore a white mask. The AI had been trained primarily on datasets dominated by white male faces, making women of color especially invisible to the system. This sparked Buolamwini’s pioneering work on algorithmic accountability, which asks not only how AI operates, but whom it serves—and whom it harms.
Beyond Neutral Technology
- Coded Bias* challenges the myth of AI neutrality by foregrounding the experiences of women of color, activists, and critical researchers. Their stories highlight how structural inequalities in society become encoded in algorithms, further embedding discriminatory practices.
- The film also raises an important question: can diversifying datasets alone solve these issues? Or do we need deeper scrutiny of the very purposes and power structures driving AI development?
Surveillance and Policing
The documentary links algorithmic discrimination to surveillance technologies, particularly in policing. While proponents of tools such as bodycams argue that they enhance transparency, research shows that surveillance is often racially skewed, targeting communities of color disproportionately.
In Finland, this debate is especially timely. Police are rolling out bodycams nationwide, with discussions about possible integration of facial recognition. Yet as U.S. and U.K. examples show, these tools can easily reinforce structural racism instead of curbing it. Moreover, as critics point out, bodycams primarily record events from the police perspective, not that of civilians.
Calls for Regulation
Civil society campaigns across the globe have raised the alarm:
- In the U.S., the Black Lives Matter movement has pressured tech companies, leading Microsoft and others to suspend sales of facial recognition to law enforcement.
- In Europe, the Reclaim Your Face campaign calls for a ban on biometric surveillance by states and corporations.
- The Council of Europe has issued guidelines for facial recognition, while the EU continues to debate AI regulation.
The UN Special Rapporteur on Racism, E. Tendayi Achiume, has stressed that discrimination is already detectable in digital technologies—sometimes intentionally, sometimes as an “unintended” disparate impact, and often in ways that compound systemic inequality.
Why This Matters for Finland
Although much of the debate has centered on the U.S., U.K., and China, the issues are equally urgent in Finland:
- Authorities are experimenting with surveillance tools without clear regulation.
- Digital literacy and critical discussion remain limited, making it harder for citizens to understand the stakes.
- Knowledge institutions—universities, libraries, schools—have a responsibility to foster awareness and demand accountability.
As Safiya Umoja Noble argues in Algorithms of Oppression, this is not just about improving technology—it’s about building multicultural and multiracial democracies that refuse to let algorithms entrench existing hierarchies.
Towards Design Justice
The principles of design justice (https://designjustice.org/read-the-principles) emphasize creating technologies that are inclusive, transparent, and equity-promoting. For this to happen, we need:
- Critical public debate on acceptable uses of data.
- Audited algorithms and clear accountability structures.
- Inclusive design practices that involve marginalized voices from the start.
As director Shalini Kantayya concludes, fighting algorithmic bias requires not only brave researchers but also engaged citizens who refuse to accept technologies that normalize inequality.
