DPE Statement of Principles on Artificial Intelligence and Professional Standards
October 5, 2026
Professionals are defined by their independent judgment: the trained, accountable, often licensed exercise of expertise on behalf of patients, students, audiences, clients, and the public. The Department for Professional Employees, AFL-CIO (DPE) embraces the AFL-CIO’s Artificial Intelligence: Principles to Protect Workers, which are the labor movement's framework for worker-centered AI.
For generations, professionals have put new technologies to work responsibly, vetting them, mastering them, and folding them into the standards of their professions, and they are prepared to do so again. What professionals cannot accept is employers deciding unilaterally how AI enters professional work, while professionals absorb the consequences: pressure to defer to systems against their own judgment, licenses on the line for a system’s errors, and the public's trust eroding when AI-shaped work fails. We must not let AI weaken professional standards. Protecting them takes both public policy and strong union contracts. The principles below are meant to guide both.
Professional judgment must remain human. AI may inform the work of professionals, but it must not replace or override the judgment of a trained or credentialed professional. A professional’s work, whether it be a diagnosis, lesson plan, news story, engineering calculation, claims adjudication, or any decision that affects the public interest, requires a human professional with the authority to accept, modify, or reject an AI system’s output.
No employer directive can put AI above a professional’s license. Professionals are personally bound by licensure, certification, and codes of ethics to standards that exist to protect their profession and the public. Employers must never require deference to an AI system where doing so conflicts with those obligations, and no professional may face discipline, retaliation, or job loss for overriding, questioning, or reporting an AI system in the course of upholding the standards of their profession.
AI must never impersonate a trained, licensed, or credentialed professional. AI systems must not represent themselves as nurses, physicians, teachers, counselors, engineers, attorneys, journalists, or any other professional, credentialed or not. Nor may AI replicate an individual professional’s voice, likeness, performance, style, or body of work without that professional’s affirmative consent, credit, compensation, and control over its use. The public has a right to know when it is receiving professional services from a person, and every professional has the right to decide whether a machine may stand in for them.
Professionals must not be held liable for AI failures. When employers require professionals to work with AI, the individual professional’s license, career, and legal standing must not be put at risk for the failures of the AI system, provided the professional exercised the judgment, diligence, and verification their profession requires. Further, no professional should be given AI-generated material to review, revise, or build upon without being told what it is.
Professionals must be trained and equipped for the AI they are required to use. Employers that deploy AI into professional work owe professionals the ongoing training to use it consistent with the standards of their professions. Professionals, through their unions, must have a role in the design, procurement, and vetting of these systems before they are deployed and in setting the standards for their appropriate use, including guarding against institutional dependence on any AI system. Leaving professionals out of these decisions risks AI that fails patients, students, and the public.
AI must not disrupt the professional career ladder. The entry-level work through which professionals learn their craft is the pipeline for every profession's future. When AI eliminates first-rung jobs such as junior reporter, staff nurse, associate, and assistant engineer, it hollows out the professions. Employers and policymakers must ensure that the path into professional careers survives the technology.
The pipeline into the professions must be tracked. Current labor data is not built to show, in near real time, what AI is doing to entry-level and early-career professional work, leaving policymakers, employers, and professionals to guess at a problem that better data could make visible. The federal government must work with states to modernize how the education-to-workforce pipeline is tracked, including by enhancing unemployment insurance wage records to connect credentials and training to specific occupations and map careers as the economy shifts. That would let policymakers see shortages, surpluses, and AI-driven hiring shifts while there is still time to respond.
AI-generated work must never masquerade as professional, human work. Anything presented as professional work that no professional authored, verified, or takes responsibility for must be disclosed as AI-generated. And no professional's name, byline, or credential may be attached to AI-generated work without their verification and consent.