Opinion from our editors and guest writers on where AI is going and where it should not.https://thefrontierwire.com/enhttps://thefrontierwire.com/icon-512.pnghttps://thefrontierwire.com/Fri, 25 Sep 2026 20:07:10 GMThttps://thefrontierwire.com/articles/stop-calling-it-hallucination/https://thefrontierwire.com/articles/stop-calling-it-hallucination/The word implies a malfunction. What the models do is closer to confident guessing, and the fix depends on naming it correctly.Sat, 12 Sep 2026 00:00:00 GMT<p>Every article about language models eventually reaches the paragraph about hallucination: the model made something up, cited a paper that does not exist, invented a court case. The word has become the industry's shorthand for being wrong. It is a bad word, and it is doing damage.</p> <h2>The word describes a malfunction that is not happening</h2> <p>A hallucination is a perception with no external cause. Applied to a language model, it suggests a system that normally reports the truth and occasionally suffers a glitch. That is not the mechanism. A language model produces the continuation that is most probable given its training and the prompt. When the probable continuation matches reality, we call it a fact. When it does not, we call it a hallucination. The model did the same thing both times.</p> <p>This matters because the word invites the wrong fix. If the problem is a glitch, you patch it. If the problem is that the system has no distinction between recall and plausible invention, you need to change what you ask of it and how you check it.</p> <blockquote> <p>The model is not lying and it is not malfunctioning. It is answering a question it was never asked: what would a good answer look like?</p> </blockquote> <h2>What to say instead</h2> <p>Confabulation is closer, and has a respectable history in neuroscience, where it describes a person filling a memory gap with a plausible story and believing it. But it still locates the problem inside the model. The more useful framing is that the model is guessing, that guessing is what it is built to do, and that the interesting question is how confident the guess is and whether anyone checked.</p> <p>That framing points at the real fixes: grounding the model in retrieved documents so the probable continuation is anchored to a source, asking for citations that can be verified mechanically, and evaluating outputs against references instead of trusting them. None of those are patches for a glitch. They are the ordinary engineering of an unreliable component.</p> <h2>Why the industry likes the word</h2> <p>It is a comfortable word. It suggests the models are almost right, that the errors are anomalies, and that a future version will not have them. It sells a roadmap. Being honest, saying the model guesses and sometimes guesses wrong, sells nothing and obliges the vendor to explain what the buyer should do about it.</p> <p>Buyers should insist on the honest version. Ask what the model's error rate is on your task, not whether it hallucinates. Ask what the product does when the guess is wrong. And when a vendor tells you the next model hallucinates less, ask them to show the evaluation. Our comparisons desk has a <a href="https://thefrontierwire.com/articles/how-to-compare-llm-evaluation-frameworks/">guide to what that evaluation should contain</a>.</p> The Frontier Wire editorsEditorialsLanguageReliabilityOpinion