06/02/2026
From the Desk of MISALC President Ron Zalewski
AI Is Not Committed to the Truth
I’m going to go off the board this month. I’ve been working on understanding AI better. The large language model and I collaborated on the following. It’s fairly decent. But I double checked every word. That’s what you’ll have to do if you want AI to be part of what you do. The generated text:
Artificial intelligence has become a common tool in workplaces, homes, and even union offices. Many representatives have experimented with AI to draft letters, summarize documents, or answer questions about labor relations. Used properly, it can save time and help organize information. Used carelessly, it can create serious problems.
The fundamental issue is simple: AI is not committed to the truth.
That statement requires some explanation. An AI system does not conduct research the way a representative researches a grievance. It does not review evidence, weigh credibility, or compare competing interpretations of contract language. Instead, it predicts what words are most likely to appear word count in response to a prompt. Often, those predictions produce useful and accurate information. Sometimes they do not.
The danger is that AI can present incorrect information in exactly the same tone and style as correct information. It may invent contract provisions, create non-existent arbitration decisions, misquote manuals, or confidently describe procedures that do not actually exist. To an inexperienced user, the answer may appear completely trustworthy.
AI can provide ideas, suggest avenues for further research, and help organize complex information. It may identify an argument worth exploring or point to issues that deserve closer examination. However, no responsible representative should file a grievance, advise a member, or challenge management based solely on AI-generated information without independently verifying the source material and confirming its accuracy.
The same rule applies to AI.
The collective bargaining agreement, the JCAM, the ELM, arbitration awards, manuals, memoranda, and established case law remain the primary sources of authority. AI can help locate issues, summarize concepts, and improve written communication. It cannot replace verification, judgment, or experience.
As AI becomes more common, our responsibility does not decrease. It increases. The representative’s job is not merely to find answers. It is to find correct answers. No matter how sophisticated the technology becomes, that responsibility remains ours.
Send a message to learn more
