Artificial Intelligence systems are inadvertently absorbing biases due to the lack of diversity in the technology sector
In an effort to increase diversity in the AI workforce and address AI bias, particularly for underrepresented groups such as women, Blacks, and Hispanics, initiatives primarily focus on workforce development, education, and training programs that emphasize inclusion. However, recent changes in U.S. federal AI policy have complicated this landscape.
Key federal efforts include the AI Workforce Research Hub and Skills Development, led by the U.S. Department of Labor (DOL), Department of Education (ED), and other agencies. These programs aim to develop AI skills across the workforce, including career and technical education (CTE), apprenticeships, and rapid retraining. The goal is to equip workers, including those from underrepresented groups, with AI competencies [1][2][3].
Industry-driven training and apprenticeships are another initiative, with a focus on creating training frameworks and expanding apprenticeships in AI-related infrastructure roles. These programs aim to build talent pipelines and provide equitable access to AI jobs by partnering with states, industries, and educational institutions [2][3].
Federal plans also include funding for rapid retraining and upskilling for individuals whose jobs may be displaced by AI. This could benefit underrepresented workers who are often disproportionately impacted by technological change [3].
However, the Trump Administration's 2025 AI Action Plan explicitly removed references to Diversity, Equity, and Inclusion (DEI) in the National Institute of Standards and Technology (NIST) AI Risk Management Framework. This removal potentially impairs efforts to embed fairness and mitigating bias within AI development from a regulatory standpoint [1].
Outside these federal initiatives, many private sector and academic organizations pursue separate efforts to reduce AI bias and promote inclusion, such as bias audits, diversifying AI datasets, and fostering diverse AI research communities.
Despite these challenges, organisations like AI4ALL are working to diversify the AI workforce by targeting historically underrepresented groups. For instance, Maya De Los Santos, an Afro-Latina woman with degrees in computer and electrical engineering, is interested in a career in AI to protect marginalized communities from AI risks and ensure they understand its benefits.
The underrepresentation of Blacks and Hispanics in the AI workforce is a significant concern. According to a Georgetown University analysis, these groups are underrepresented, and approximately 60% of public high schools offer AI-related classes, with underrepresented groups less likely to have access [4].
Moreover, men hold 80% of tenured faculty positions at university AI departments globally, and among AI technical occupations, Hispanics hold about 9% of jobs, compared with holding more than 18% of US jobs overall. Black workers hold about 8% of the technical AI jobs, compared with holding nearly 12% of US jobs overall [5].
The slow progress in increasing the representation of women in AI is also evident. According to UNESCO, from 2021 to 2024, the number of women working in AI globally increased by only 4%, and women represent 26% of the AI workforce [6].
Safiya Noble, a professor at the University of California Los Angeles, worries that the government's attack on DEI will undermine efforts to create opportunities in AI for marginalized groups. Noble argues that the government's backlash against movements like Black Lives Matter and allegations of anti-conservative bias are evidence of this repression [7].
Despite these challenges, Tess Posner, CEO of AI4ALL, remains optimistic about the commitment to values of inclusion in the AI field [8].
References: [1] https://www.nist.gov/itl/applied-cybersecurity/ai-risk-management-framework [2] https://www.whitehouse.gov/artificial-intelligence/ [3] https://www.whitehouse.gov/wp-content/uploads/2019/02/Executive-Order-13859---Maintaining-American-Leadership-in-Artificial-Intelligence.pdf [4] https://www.georgetown.edu/sites/default/files/report_ai-and-underrepresented-minorities-in-us-education.pdf [5] https://www.brookings.edu/research/the-underrepresentation-of-women-and-minorities-in-ai-careers/ [6] https://en.unesco.org/themes/information-communication/artificial-intelligence/women-in-ai [7] https://www.wired.com/story/government-backlash-against-diversity-equity-and-inclusion-threatens-ai/ [8] https://www.ai4all.org/
- The AI4ALL organization is making strides to diversify the AI workforce, targeting historically underrepresented groups like women, Blacks, and Hispanics, who often face barriers in accessing AI education and jobs.
- The removal of Diversity, Equity, and Inclusion (DEI) references from the National Institute of Standards and Technology (NIST) AI Risk Management Framework by the Trump Administration is a potential impediment to the integration of fairness and bias mitigation within AI development.
- Addressing climate change and reducing carbon emissions through AI techniques is an untapped area of collaboration between the business, finance, and technology sectors, especially considering the power of AI in optimizing energy usage and promoting sustainable practices.
