As corporations increasingly embrace artificial intelligence (AI) to streamline operations, diversity, equity, and inclusion (DEI) training has become one of the latest frontiers for the technology. AI-powered platforms promise scalability, consistency, and cost efficiency, but deploying them in place of in-person DEI training carries risks that the programs will seem irrelevant and do little to change individual or organizational behavior. They could easily be counterproductive in the eyes of those most likely to experience unconscious bias and hidden forms of discrimination.
RISKY BUSINESS
Here are some dangers of using AI as the single medium to provide DEI training:
• Illusion of neutrality. One of the most dangerous assumptions about AI is that it is inherently objective. AI systems are based on historical data that often reflects systemic biases. When applied to DEI training, these biases can be reproduced and even amplified. An AI module designed to teach employees about unconscious bias may itself operate from a data set that underrepresents certain communities, skews cultural perspectives, or defaults to a dominant worldview. Rather than dismantling prejudice, such a tool risks quietly reinforcing it, while projecting an air of scientific credibility that discourages scrutiny.
• Loss of human connection and psychological safety. DEI training is not a knowledge transfer exercise; it is a deeply human process. Its effectiveness depends on vulnerability, dialogue, and the gradual building of trust between participants and facilitators. In-person training creates an environment where employees can ask sensitive questions, share personal experiences, and engage in the kind of honest, sometimes uncomfortable conversations that drive genuine cultural change.
In some of my most memorable DEI programs, participants would challenge my “biased” perspective and terminology. One such example was when I was referring to the bias in intelligence testing, which was based on dominant group experiences. To counter this bias, diversity scholars came up with an equally biased intelligence test based on the black American experience where whites would score lower than blacks. The scholars called this the “Chitlin’s Test.” In one program, black participants took strong exception to my using the name of the test because of its underlying reference to what slaves were provided to eat. This led to my appreciation of their perspective and an engaging discussion by the whole class regarding the use of words that have hidden biases and triggers. AI platforms, regardless of how sophisticated they become, cannot replicate this dynamic. A chatbot cannot read the room, sense when someone is distressed, or respond with empathy to an employee who discloses a personal experience of discrimination.
• Customization that addresses individual and organizational issues. Programs can only be relevant and effective if they address the key issues facing the participants. This customization is based on interviews, focus groups, and individual anonymous and confidential pre-program questionnaires that capture the real lived experience of each participant. These can be turned into case studies for the participants to address. Each participant also should fill out their own individual and organizational action plans, which the trainer can consolidate and then share with the participants, their managers, and organizational leadership.
• One-size-fits-all content in a diverse world. Effective DEI training must be context sensitive. The cultural dynamics of a tech startup in California differ vastly from those of a manufacturing plant in rural Georgia or an international office in Singapore. In-person facilitators can tailor content, language, and examples to the specific makeup of their audience in real time.
AI systems tend to deliver standardized content. While customization features exist, they are often superficial. An AI that presents the same scenarios, the same language, and the same cultural assumptions to every employee across every geography is inherently not inclusive by its design. This homogenization can alienate the very employees the training is meant to reach. Many U.S.-based DEI training programs fail outside the U.S. because they only use U.S. examples or priorities while ignoring local issues. While race may be a critical issue in the U.S., it may have little relevance in other countries. In our training in Saudi Arabia, one of the big issues was bias based on degrees of religiosity, while in China, it was urban vs. rural upbringing.
• Accountability and measurement gaps. In-person DEI training allows organizations to measure genuine engagement: Facilitators observe body language, moderate discussions, and provide nuanced feedback to leadership. AI platforms generate data, completion rates, quiz scores, and time spent on modules, but these metrics say little about whether attitudes have shifted or behaviors have changed or if organizational policy biases were addressed.
There is also a risk that corporations use AI-based DEI training as a liability shield rather than a tool for real transformation. Documenting that every employee “completed” an automated module is very different from demonstrating a sustained commitment to equity. When DEI training is driven by compliance rather than conviction, AI makes it easier than ever to appear committed while doing the minimum.
IS THERE A PLACE FOR AI IN DEI TRAINING?
AI can play a supporting role in DEI training, delivering foundational content, reinforcing concepts between training sessions, and/or providing on-demand resources to employees. However, it should supplement, not replace, the irreplaceable human element of in-person facilitation.
Corporations that treat AI as a cost-saving substitute for meaningful DEI training risk doing real harm to their culture, to marginalized employees, and ultimately to the long-term organizational health that genuine inclusion builds. The most forward-thinking companies will use AI as a tool in service of human connection, not as a replacement for it. AI as the primary vehicle for teaching about diversity, equity, and inclusion can become counter-productive and damage an organization’s underlying goals and reputation.
If you have any questions or cases where AI was used exclusively to deliver DEI training, please send them to me at: Neal@NealGoodmanGroup.com