Artificial intelligence that runs on algorithms has been integrated into nearly every aspect of life. But what if bias is embedded in the coding of those algorithms? Coded Bias begins with PhD student Joy Buolamwini's discovery that facial recognition software couldn’t recognize her face until she donned a white mask. Joy began examining the data sets used to train AI and discovered the data was overwhelmingly comprised of white male faces, making the tech unable to correctly identify darker – particularly darker female – faces.
Director Shalini Kantayya's documentary loosely follows Joy as she sheds light on this issue through her research. Everything from resume screening to health insurance rates are frequently determined by algorithms that have been found to replicate racial and gender bias.
The alarming implications of algorithmic bias are far-reaching and effectively showcased here through both research and impactful personal stories. Biased algorithms have the potential to cause tangible harm as seen by the parolee who was penalized after an algorithm determined she was "high risk" despite her exemplary behavior and the 14-year-old Black boy in the U.K. who was grabbed off the street and frisked by police after being misidentified through facial recognition software.
Although algorithmic bias is the jumping-off point, the documentary shifts focus to global AI usage concerns. With the advent of facial recognition technology, big-brother-style mass surveillance is now a reality. Beyond inherent privacy concerns, the technology's inaccuracy for some facial types makes its widespread usage even more troubling. The experts consulted throughout the film make a strong case for the need for legal regulations to prevent both the State-style surveillance already seen in China, as well as corporate use of private personal data.
Coded Bias brings to the fore an important discussion that has implications for people across the globe. The documentary's cautionary message about the dangers of algorithmic bias is presented in an engaging and humanistic way. Technical details are kept to a minimum, which could be viewed as positive or negative depending on prior knowledge and interest.
Instead, the focus is on the social implications of ever-present biased technology in a world that increasingly relies on its supposed objectivity. Joy calls algorithmic justice one of the biggest civil rights issues of our time, and with all the ways algorithms impact daily life, it's not hard to see why. Recommended. Aud: C, P.