Coded Bias: How AI can perpetuate racial stereotypes

Описание: The inspiration for this project came from the growing discourse of algorithmic bias in AI I have witnessed firsthand or in the media. Although I believe AI can enhance your efficiency and decision-making, there’s also a complex issue that is not often talked about enough. In Coded Bias, I delve into the exploration of racial bias in AI through the medium of video art. Through a critical examination of existing literature, news articles, and other data scientists/activists, I aim to shed light on the complexities of this issue and provoke thought on the implications for society.

I began my project by doing research on ways racial bias affects our society today. Studies have revealed these biases manifest in various domains, including criminal justice, healthcare, employment, and social media, reinforcing existing systemic injustices. I drew a lot of inspiration from Joy Buolamwini, a poet of code who uses art and research to illuminate the social implications of artificial intelligence.

Through visual imagery, soundscapes, and narrative storytelling, the piece juxtaposes the promise of technological advancement with the reality of systemic discrimination. Drawing inspiration from real-world examples and a poem written by Joy Buolamwini titled AI, Ain’t I A Woman. Most of the footage used in the video is found. I initially had the idea of using my own screen recordings using these AI platforms, but most of the shows in the video were inaccessible to the public, and the ones that I tried using weren’t as interesting to use. I tried to add some twists to the found footage by playing with the audio and adding layovers as well as transitions to make them seamlessly blend.

Creating this video art piece has prompted me to confront my own biases and assumptions about technology and society. Through the process of research and artistic expression, I have gained a deeper appreciation for the complexity of the issue and the importance of interdisciplinary dialogue. Many activists state that having more people of color in the AI engineering and computer science spaces can help combat these biases. Moving forward, it is essential to continue advocating for transparency, accountability, and diversity in AI development to mitigate the harmful impacts of bias and advance social justice.

Sources

NPR- lack of diversity in AI fields:
https://www.npr.org/2022/02/13/1080464162/...

Google Photos Scandal:
https://www.bbc.com/news/technology-33347866

Video of black women being put into AI generators:
https://time.com/5520558/artificial-intell...

Facial Recognition AI:
https://www.faceplusplus.com/