Artificial intelligence is no longer a future-of-work topic; it is already showing up in recruiting, résumé screening, video interviews, workplace surveillance, training recommendations, pay decisions, and even layoffs or termination decisions. That matters in Minnesota because the legal risk is not only whether a tool is “smart,” but whether it changes outcomes in a way that is discriminatory, unreviewed, or impossible to explain. (eeoc.gov)
For Minnesota employers, the starting point is the Minnesota Human Rights Act. The Minnesota Department of Human Rights says employers may not discriminate because of race, disability, sex, religion, national origin, age, sexual orientation, gender identity, and other protected identities, and may not punish an employee for reporting discrimination. As of October 1, 2025, Minnesotans who want to preserve any related federal rights must separately and independently file with the U.S. Equal Employment Opportunity Commission. (mn.gov)
Federal law adds another layer. The EEOC says that neutral policies and practices can still violate the law if they have a disproportionately negative effect on protected groups and are not job-related and necessary to the business. The agency also warns that AI tools can unintentionally screen out people with disabilities, including in résumé review, video interviews, and other automated assessments. In other words, if an algorithm is making the first cut, the employer still owns the decision. (eeoc.gov)
What employers should do now
The safest approach is to treat AI like any other employment tool that needs supervision, documentation, and a clear business purpose. Start by making an inventory: where is AI used, who chose it, what data does it analyze, and which employment decisions can it influence? That inventory should cover recruiting, applicant ranking, interview scoring, scheduling, performance management, promotion, discipline, pay, and separation decisions, because those are the places the EEOC specifically flags as risk points. (eeoc.gov)
Next, build human review into the process. That is partly a legal defense and partly common sense: if a tool is being used to narrow the candidate pool or rank employees, a manager should be able to review the results, question outliers, and override the system when the facts do not fit the output. If a vendor says the tool is “bias free,” ask for evidence, not branding. The federal guidance on disparate impact makes clear that a neutral-looking tool can still create unlawful results if it is not actually job-related and necessary. (eeoc.gov)
Employers should also think carefully about accommodations. The EEOC notes that even when AI is involved, existing laws may still require reasonable accommodation based on disability, religion, or pregnancy-related limitations. That means a standard online assessment may need a different format, extra time, a human alternative, or a modified workflow. If the company uses video, voice, keystroke, or location-based analytics, the employer should be ready to explain why those signals are needed and how the company avoids unfairly penalizing workers with disabilities. (eeoc.gov)
Finally, make the process transparent. At minimum, employees and applicants should know when AI is being used in a way that affects them. Transparency is not just a public-relations issue; it helps workers request accommodations, challenge errors, and understand what data is being used to judge them. It also gives employers a chance to correct the output before a bad recommendation becomes a bad hiring or discipline decision. (eeoc.gov)
What workers and applicants should watch for
If you are applying for a job or trying to advance in your current one, look for signs that software is quietly doing the deciding. Common examples include job ads that change based on your profile, résumé filters that reject you for missing keywords, chatbot screeners, recorded video interview analysis, or systems that track your location, keystrokes, cursor movement, or customer feedback. Those are all examples the EEOC says can show up in the AI workplace. (eeoc.gov)
If something feels off, document it. Save the job posting, the assessment link, messages from the employer, and any feedback you receive. If you need an accommodation, ask for it clearly and early. If you believe the AI tool is creating a discriminatory outcome, Minnesota workers can raise concerns under state law, and may need to file separately with the EEOC to protect federal rights. (mn.gov)
A likely next step in Minnesota policy
Minnesota lawmakers have already introduced bills in 2026 that would regulate the use of artificial intelligence in employment settings, including provisions aimed at discrimination and notice. Those bills are not the same as enacted law, but they signal where the conversation is headed: more transparency, more accountability, and more attention to how automated systems affect real people. Employers that get ahead of those themes now will have less scrambling later. (revisor.mn.gov)
The bottom line is simple. AI can help a Minnesota workplace move faster, but it cannot replace the obligation to hire, supervise, promote, and discipline fairly. Employers should map their systems, test for bias, keep humans in the loop, and make accommodations workable. Workers should stay alert, keep records, and speak up early when an automated tool starts shaping a human decision. In Minnesota, that combination of transparency, documentation, and accountability is the best protection both sides can have. (mn.gov)