Black Pride Day invites us to reflect on the importance of the Black community in Brazil, but also on the complex relationship between technology and racial issues. As an innovation that is increasingly impacting our lives, artificial intelligence (AI) has been reflecting biases rooted in the algorithms that drive it.
In 2019, a study conducted by the National Institute of Standards and Technology (NIST) in the United States showed that facial recognition systems had error rates up to 100 times higher for Black and Asian individuals compared to white individuals.
Coded Bias Documentary
A striking example is presented by the project of MIT Media Lab graduate student Joy Buolamwini, called Aspire Mirror. When trying to create an inspiring mirror, the researcher encounters problems when the AI fails to recognize the faces of women and people with darker skin tones.
In the documentary Coded Bias, shown on Netflix, Buolamwini explains that she was working on the concept of “computer vision” by merging images of a face in front of a mirror and generating a reflection based on inspiration—like a creative prompt. However, the camera wouldn’t recognize her face.
After several tests, the Ghanaian-American scientist decides to wear a white mask, and only then does the system respond. From that moment forward, she commits to investigating why both hardware and software consistently fail to identify people with dark skin and show a strong tendency to recognize white men.
In her research, Buolamwini tested more than 1,200 facial recognition images from major technology companies such as Google, Microsoft, and Amazon. The results confirm that this type of software repeatedly fails to recognize Black individuals because the datasets used to train the AI are underrepresentative of diversity, resulting in automated bias and reinforcing stereotypes and discrimination.
Biased Algorithms
The data used by AI usually comes from historical information and is applied to make predictions. In this way, artificial intelligence ends up replicating the social world and reproducing inequalities.
This does not occur because the technology lacks ethical criteria, but because it is a mathematical process of information handling that relies on automatically detecting patterns present in datasets. This reflects racism as a historical data base, constructed from flawed, incomplete, or generalized information—as shown in the documentary.
Thus, these findings reveal the harmful influence of biased data in training these algorithms, and while AI brings innovation, it is not immune to replicating and perpetuating racial inequalities.
As presented in the documentary, the solution lies in strengthening legislation and creating oversight organizations capable of enforcing protections that defend civil rights.
Artificial intelligence is rapidly becoming integrated into business, and awareness of the importance of its ethical use is also growing. That is why Black Consciousness Day reminds us that the fight against racism also involves our human responsibility to ensure that technology is truly inclusive and respects equality among all people.
