How AI can advance Diversity, Equity, and Inclusion in Social Work Education and Research
The advancement of AI technology could offer unprecedented opportunities to advance diversity, equity, and inclusion (DEI) in social work education and practice. While misuse of AI can deepen existing inequities, its thoughtful application has the potential to empower underserved groups, democratize access to resources, and reshape how we address social issues. In this post, I will discuss several opportunities to advance DEI in social work education and research by harnessing AI.
Democratizing Data Analysis in Education
Quantitative data analysis, historically considered a costly and specialized skill, has been out of reach for many nonprofits and organizations serving marginalized communities. Over the past three years, as a Community Research Consultant for NYU Langone Hospital, I’ve supported nonprofits, particularly those working with immigrant populations, in structuring and analyzing data for grants and needs assessments. The challenge lies in the scarcity of quantitative data focused on minority or immigrant populations—data crucial for advocacy and policymaking.
On the other hand, I’ve seen students with no technical background use AI tools to analyze data and solve problems, gaining skills and confidence that lead to better opportunities. By training communities on how to use these tools effectively, we can turn AI into a vehicle for inclusion and empowerment. AI tools such as ChatGPT are revolutionizing how we approach data analysis, making these skills more accessible and enabling organizations and individuals to conduct advanced analyses without formal training. Green (2024), building on the work of Chine et al. (2022) and van der Vorst & Jelicic (2019), highlights how AI platforms lower barriers to entry among Black, Latinx, and Afro-Caribbean communities by providing personalized feedback, automating repetitive tasks, and offering adaptive learning experiences. These advancements are particularly relevant for social work education, where the ability to analyze and interpret data is critical for needs assessments, grant writing, and policy advocacy.
Here is a quote from my working paper on data science education with social work PhD students: “There is now a platform that offers free or low-cost access to these (data science) methods, depending on the model you choose. Importantly, the rigor of the learning process has not been compromised. (…) This platform makes it possible to execute these methods much faster than before, leveling the playing field. Previously, only full-fledged professors with the financial means and access to extensive Python classes and mentors versed in NLP could pursue these methods. Now, anyone with an interest can learn and apply them from the comfort of their home, making these approaches more attainable.”
Salas-Pilco et al. (2022) discuss how AI-powered tools promote inclusive education by addressing technological and pedagogical barriers. For example, consider a social work research methods class focusing on quantitative methods. Traditionally, students from underfunded schools or nontraditional backgrounds may struggle due to limited access to advanced statistical software or training in data science. AI tools can bridge this gap by providing real-time guidance, suggesting data visualization techniques, and simplifying statistical operations. Students can use AI to clean datasets, conduct regression analyses, or create policy-relevant visualizations, significantly enhancing their learning experience.
However, while AI holds promise, it is not without challenges. Privacy concerns and potential biases in training data require careful consideration. As van der Vorst & Jelicic (2019) and Chine et al. (2022) emphasize, responsible AI use is essential to ensure that democratization does not inadvertently reinforce existing inequities. Social work educators need to teach students not only how to use these tools but also how to critically evaluate their outputs and limitations based on the NASW Standards for Technology in Social Work Practice.
Addressing Research Gaps
AI tools can empower historically underrepresented groups by providing accessible opportunities to engage in data-driven decision-making and social innovation. Research has historically underrepresented minority populations via structural barriers, such as limited funding and resources. My own work illustrates how data science can fill these gaps. One of my studies analyzed the nationwide distribution of mental health service language accessibility, highlighting disparities for non-English-speaking populations. Historically, research on non-English speakers has been underrepresented and costly to conduct. AI tools have allowed me to quantify and visualize these disparities, creating data-driven insights for advocacy.
W. E. B. Du Bois’s Data Portraits: Visualizing Black America serves as a historic example of how data visualization and mapping can be powerful tools for minority and racial studies, as well as for advocacy. His pioneering work demonstrates how quantitative data visualization can bring attention to social inequities and serve as a medium for justice and reform. Similarly, contemporary quantitative data visualization and analysis can be leveraged to promote social justice, diversity, equity, and inclusion. However, proficiency in such quantitative methods has long been perceived as exclusive to STEM disciplines, further perpetuating barriers for many underrepresented communities. AI tools can help bridge this gap by making these methods more accessible and enabling historically marginalized groups to harness data for meaningful advocacy and change.
Moreover, digital data offers unique opportunities for understanding marginalized groups. Platforms like Black Twitter and online queer support groups reveal how individuals build community and seek support in ways that traditional research often overlooks. These digital traces provide a less intrusive way to study needs, perceptions, and risk or protective factors, ultimately contributing to equity in research and practice.
Enhancing Language Accessibility
Language accessibility is a critical equity issue, especially for immigrant and limited-English-proficiency (LEP) populations. Recent advancements in natural language processing (NLP), including large language models (LLMs), have significantly improved machine translation and language interpretation technologies. Public agencies in states like New York and New Jersey already use tools like Google Translate to make their websites accessible to LEP users. Such tools have become a daily necessity for many non-English speakers, even in healthcare settings like patient portals (Rodriguez et al., 2021).
Research on AI-driven language interpretation shows potential for improving healthcare and public service accessibility (Bakdash et al., 2024). While current technologies still face issues of accuracy and trust, ongoing advancements, especially in low-resource languages, offer promising solutions. For example, Google recently added an Indigenous Canadian language to its translation service, marking a milestone in addressing linguistic inequities (See here). Such efforts demonstrate the potential of AI to enhance accessibility for linguistic minorities and immigrant communities, who often face significant barriers to information and services.
Responsible Use for Equity
While AI offers transformative opportunities, its ability to advance DEI depends on how responsibly it is implemented. Ethical challenges such as data privacy, bias, environmental cost, and English hegemony must be addressed. Companies like Meta, Google, and OpenAI are now investing in improving low-resource language models, aiming to avoid exacerbating English hegemony while expanding linguistic diversity in AI (World Economic Forum, 2024). My work focuses on ensuring AI is used to empower rather than exclude. By equipping students and practitioners with the knowledge and skills to use AI tools ethically, we can ensure these technologies become a force for equity, expanding opportunities for all.
