REPORT: Universities Push AI Tools As Trojan Horse for DEI Agenda

The parental rights watchdog Defending Education published a new report last week warning that prominent universities are working to integrate diversity, equity, and inclusion (DEI) priorities into artificial intelligence (AI) programs that could eventually shape how AI is used in K-12 classrooms.

The basic argument is straightforward. AI is already becoming a major part of American education, helping teachers develop lesson plans, assisting students with research, providing personalized tutoring, and even evaluating student work.

Defending Education argues that activist-minded university professors understand this growing trend and are working now to make concepts like “equity,” “social justice,” and “algorithmic bias” central to the way AI is developed, evaluated, and taught.

If those principles become accepted as standard components of “responsible AI,” the report warns, DEI ideology could effectively become baked into the educational tools used by millions of students.

“The integration of AI into education is not a matter of if but when,” Defending Education states. The organization argues that universities are therefore attempting to “institutionalize DEI” in AI while the technology and the rules governing its use in schools are still taking shape.

Casey Ryan, an investigative reporter at Defending Education, said universities did not truly reverse course when the Trump administration threatened their funding for promoting DEI initiatives and programs last year. Instead, he argues, many simply found new avenues for promoting the same ideas.

“In an attempt to save face last year, many universities pretended to backtrack in public-facing actions and statements from the DEI priorities and programs that they had supported and funded for years at the taxpayer expense,” Ryan told The Center Square. “We can now see that, in many cases, it was mostly a thinly veiled attempt to protect their own funding from the federal government. Our new report exposes how institutions of higher learning didn’t really retreat from DEI; they just found a new way to promote it under the radar.”

One of the most significant examples highlighted in the report is the Massachusetts Institute of Technology’s (MIT) R”esponsible AI for Social Empowerment and Education initiative,” or RAISE. MIT describes RAISE as its “home for AI education” and says its mission is to create a more “inclusive future of education and learning in the era of AI.”

That effort already has considerable reach. MIT says its “Day of AI” curriculum has reached 1.5 million students and more than 50,000 teachers and has been used in all 50 states. RAISE also offers free AI curricula for K-12 students, including a middle-school ethics curriculum that teaches about “algorithmic bias.”

One exercise intentionally gives students a biased dataset to train an AI system, allowing them to observe how faulty or unrepresentative data can produce skewed results.

Defending Education does not object to teaching students that AI can produce inaccurate or biased results. Rather, the report argues that universities are increasingly defining “bias” through an ideological framework involving race, gender, language, and other identity categories and then treating “equity” as a necessary corrective for bias.

In other words, America’s academic elites are taking a two-pronged approach to AI. First, they are training educators to teach students that any AI-drive response that conflicts with DEI principles must be “biased.” Second, they are attempting to manipulate AI models to only produce responses that align with the DEI agenda.

The University of California, Berkeley provides another example. Its Responsible AI Initiative says its work encompasses “fairness, equity, privacy, safety, security, and accountability” and seeks a more “inclusive and just world.”

One Berkeley project questions whether ChatGPT discriminates against speakers because it defaults to “Standard American English,” as opposed to, for example, so-called “Black English.” Berkeley Researchers contend that discrimination based on language can serve as a proxy for discrimination based on race, ethnicity, or nationality.

Another project, “Gender Equity & Generative AI,” describes generative AI systems as “stereotype machines” and seeks ways to identify and mitigate perceived “gender bias.”

At Columbia University, researchers have gone even further in developing tools for measuring AI through an explicitly equity-focused lens. Columbia’s Social Intervention Group operates an “AI for Social Good and Society” initiative that envisions AI “rooted in equity and community engagement.”

Among Columbia’s projects is “Human-Centered Benchmarks for Evaluating AI Chatbot Equity,” which evaluates chatbot responses for LGBTQ+ communities. Another project studies how prompts can be structured to make responses from ChatGPT, Claude, and Gemini more inclusive, including a test case involving mental-health guidance for “queer BIPOC youth.”

Morgan State University similarly operates a “Center for Equitable Artificial Intelligence and Machine Learning Systems,” or CEAMLS, whose mission includes facilitating “socially responsible and equitable artificial intelligence systems.” The center says it develops standards and testing protocols intended to identify and mitigate “algorithmic bias.”

Importantly, CEAMLS’s work is not confined to university researchers. Morgan State runs a K-12 AI education program and in 2026 offered teachers a six-week summer training cohort, complete with a $4,000 stipend, to develop classroom-ready AI resources. The center also says it has launched outreach programs engaging students in “responsible AI.”

Its published research includes papers on racial and gender bias, “culturally relevant” language models, universal basic income, and one document conspicuously titled, “What is the Future of Work in the Generative AI Era? A Marxist and Ricardian Analysis.” The center has also hosted discussions on “Marxist perspectives on AI.”

Defending Education’s concern is broader than a handful of politically charged college courses. Universities are training teachers, educating students, conducting research, designing evaluation standards, advising policymakers, and collaborating with technology companies at the precise moment when schools are beginning to determine how AI will be used in classrooms.

The report argues that whoever establishes the standards for “responsible” and “ethical” AI today could exercise enormous influence over what those terms mean in education tomorrow. Much like far-left ideologues captured education curricula in decades past, the political left is now seeking to ensure that any AI tools deployed in classrooms are just as “woke.”

For parents who believed the retreat from formal DEI offices and programs meant the ideological battle in education was ending, Defending Education’s findings suggest that this battle may instead be moving into a new and potentially far more consequential phase.