AI hallucinations
Also used: Hallucination, AI hallucination
An AI hallucination is a response that sounds plausible but includes information that is invented, unsupported, or inconsistent with the available source material.
Why it matters to educators
Educators make decisions about learning, evaluation, policy, and communication. A confident tone cannot be treated as evidence, especially when an AI response names a source, interprets a rule, or makes a claim about a student.
The foundation
Generative models predict a plausible response; they do not automatically verify every claim before presenting it. When the prompt lacks reliable source material—or when the model cannot distinguish a strong pattern from a true fact—it may fill a gap with something convincing but false.
Good prompts can reduce some errors but cannot eliminate hallucinations. For consequential work, educators should provide authoritative sources, require traceable evidence, verify important claims independently, and avoid delegating final decisions to the model.
What this can look like in education
A citation that does not exist
A chatbot generates a polished literature summary with an invented journal article. The educator verifies each citation in a trusted database before using or sharing the summary.
A confident policy error
An AI assistant states the wrong deadline from a district policy. The administrator checks the claim against the current policy document before communicating it.