Safety
Hallucination
When an AI confidently states something that is factually wrong or completely made up.
01 ——
In plain English
Hallucination is when a language model generates text that sounds plausible and confident but is factually incorrect. The model isn't lying — it doesn't have intentions — it's pattern-matching on training data and sometimes produces outputs that have the form of facts without the accuracy.
Common examples:
- Fabricating citations, book titles, or URLs that don't exist
- Confidently giving the wrong date, statistic, or name
- Inventing legal cases, scientific studies, or product features
Why it happens: LLMs are trained to produce likely text, not true text. They don't have a built-in fact-checker.
How to reduce it: Retrieval-Augmented Generation (RAG), asking the model to cite sources, and human review for high-stakes outputs.
02 ——
Related terms
LLM
Large Language Model — the type of AI behind tools like ChatGPT and Claude, trained to understand and generate text.
RAG
Retrieval-Augmented Generation — a technique that gives an AI model access to external documents before it answers, so it can cite real, up-to-date sources.
Prompt Injection
A security attack where malicious instructions hidden in user input or external content trick an AI model into ignoring its real instructions.
Alignment
The challenge of making AI systems behave in ways that match human values and intentions — not just their literal instructions.
Guardrails
Rules and filters that constrain what an AI model can output — used to block harmful, off-topic, or non-compliant responses.
Back to glossaryLast reviewed May 2026