Article

Beyond the Hype: Exploring AI's Dark Side and Its Anti-Black Biases

Published
September 5, 2023
Marysabel Villafuerte

The emergence of artificial intelligence (AI) and its potential, particularly exemplified by platforms like ChatGPT, has generated widespread interest. However, along with the excitement,there is growing concern about the dark side of AI—its inherent anti-black bias. This bias manifests in various AI applications and systems, causing harm and perpetuating racial inequalities.

One area affected is recruiting and hiring platforms. A 2021 report by the University of Pennsylvania Law school Policy Law1 investigating bias in these systems found that AI tends to favor keywords and criteria commonly associated with white applicants, disadvantaging Black candidates. For instance, historical biases could lead AI to devalue resumes mentioning HBCUs, potentially disqualifying well-qualified individuals. The phenomenon of stereotype threat, where individuals feel stereotyped based on their unique identities, further complicates the matter, leading to a lack of inclusivity in recruitment processes.

Facial recognition technology is another area rife with anti-blackness. Numerous wrongful arrests have occurred due to facial recognition software falsely identifying innocent Black individuals as criminals. Nijeer Parks was sent to jail due to a false facial recognition, he spent 11 days in jail2, and Robert Williams was arrested and spent 30 hours in a detention center3. This technology's erroneous and biased outcomes have disproportionately impacted Black men, highlighting the grave consequences of unchecked AI bias.

The realm of social media filters perpetuates white beauty standards, reinforcing colorism and distorting perceptions of racial diversity. These filters often lighten skin tones and alter facial features, promoting an unrealistic and biased portrayal of beauty, favoring white traits over others.

AI-powered photo generation apps exhibit similar problems. Users have reported frustration with platforms smoothing skin, brightening features, and even erasing unique racial characteristics. These apps often disregard or distort traits associated with Black identities, further contributing to the problem of anti-blackness in AI.

Even generative AI platforms like ChatGPT are not immune to biases. Studies have shown that given certain prompts, these systems can produce overtly racist results. Their responses might reflect harmful stereotypes and opinions, underscoring the need to address and mitigate AI's inherent biases.

Addressing these issues requires a concerted effort to infuse AI development with an anti-racist and anti-oppressive approach. Currently, the lack of diversity in AI development teams exacerbates these biases, necessitating the establishment of AI ethics committees and increasing representation of Black individuals and other marginalized communities in the field.

In conclusion, while AI brings forth tremendous possibilities, it also poses significant risks due to its anti-black bias. To create a more equitable future, it is essential to be aware of and challenge the biases ingrained in AI systems and strive for inclusive and ethical development practices. To read the article written by Janice Gassam, click in the following link: https://n9.cl/g952q

1.- https://www.law.upenn.edu/live/files/11567-the-elephant-in-ai

2.- A false facial recognitionmatch sent this innocent black man to jail (https://edition.cnn.com/2021/04/29/tech/nijeer-parks-facial-recognition-police-arrest/index.html)

3.-Detroit man sues police forwrongfully arresting him based on facial recognition (https://www.theverge.com/2021/4/13/22382398/robert-williams-detroit-police-department-aclu-lawsuit-facial-recognition-wrongful-arrest)

Check out our HCI case studies here.
Read next
Singapore's Investment in AI: An Ambitious Path, Yet with Pending Challenges

Innovation with Responsibility: The European Union Leads the Way in AI Regulation

Navigating the Future: Analyzing the Biden Administration's Strategic Approach to Responsible AI Development.