Artificial intelligence systems are being quietly influenced by the lack of diversity in the tech industry, leading to built-in biases.
In the realm of Artificial Intelligence (AI), experts and observers have noted a concerning dominance by white men. Despite this, there are efforts to make AI more inclusive, with Maya De Los Santos remaining optimistic about these endeavours.
Recent reports from UNESCO reveal that the number of women working in AI has increased by only 4% from 2021 to 2024. This statistic underscores the slow progress towards a more diverse AI workforce.
The underrepresentation of minorities, particularly blacks and Hispanics, in the AI sector is a significant concern. For instance, black workers hold about 8% of the technical AI jobs, compared with nearly 12% of US jobs overall. Similarly, among AI technical occupations, Hispanics hold about 9%, compared with more than 18% of US jobs overall.
The lack of diversity in AI can lead to biased outcomes, posing risks in various sectors such as hiring processes, healthcare, finance, and law enforcement. For example, Amazon.com scrapped an AI recruiting tool when it was discovered that the system was favouring male candidates over female ones.
Maya De Los Santos, an Afro-Latina woman with degrees in computer and electrical engineering, is one of many aspiring to forge a career in AI. However, efforts to make AI more representative of American society are facing challenges due to the backlash against Diversity, Equity & Inclusion (DEI) programs.
Safiya Noble, a professor at the University of California Los Angeles, shares concerns that this backlash against DEI will undermine efforts to create opportunities in AI for marginalized groups. She warns that the government's attack on DEI could hinder the growth of AI for the benefit of all.
Efforts to address these issues globally include various initiatives, partnerships, and educational programs aimed at workforce development, ethical AI practices, and bias reduction. These measures range from workforce development and retraining programs to ethical AI guidelines and frameworks, diversity and bias mitigation efforts in AI development, and industry and education sector partnerships.
However, recent major U.S. policy directions, such as the 2025 America's AI Action Plan released by the Trump Administration, focus primarily on AI dominance and rapid adoption with less emphasis on DEI. This shift towards deregulation and rapid market adoption could potentially exacerbate the existing disparities in the AI workforce.
Access to AI education is also a concern, with about 60% of public high schools offering AI classes, but underrepresented groups like Blacks, Hispanics, and Native Americans less likely to have access.
In conclusion, while many global stakeholders are promoting diversity in AI through education, ethical standards, and workforce initiatives, the recent prominent U.S. federal policy shifts prioritize AI innovation and infrastructure over explicit DEI goals. This could have significant implications for the future of AI and its impact on society.
- The lack of diversity in AI workforce, as reflected by UNESCO's report, indicates a slow progress towards inclusivity, with women accounting for only 4% more of AI jobs from 2021 to 2024.
- Underrepresentation of minorities in the AI sector is pronounced, with black and Hispanic workers comprising 8% and 9% of technical AI jobs, respectively, a stark contrast to their representation in US jobs overall.
- These disparities can lead to biased outcomes in sectors like hiring, healthcare, finance, and law enforcement, as demonstrated by Amazon's discarded AI recruiting tool, which favored male candidates.
- Safiya Noble, a professor at the University of California Los Angeles, expresses concern that the backlash against Diversity, Equity & Inclusion (DEI) programs could impede opportunities for marginalized groups in AI and thus, hamper the growth of AI for the benefit of all.
