Hallucination
Confident outputs that are factually wrong.
"Confidence and accuracy are different variables. Models can max either."
What is Hallucination? Confident outputs that are factually wrong. Major source of enterprise AI risk.
LLMs inventing citations, case law, or factual claims.
Major source of enterprise AI risk.
Retrieval grounding. Citation requirements. User training.
An LLM gives a confident citation: 'Smith et al., 2019, Journal of Finance, p.142.' You check. The paper doesn't exist. Why?
Pick a reaction to Hallucination
One tap. We'll point you at the most useful next surface based on how this hits.
The full taxonomy entry
Every concept in the Atlas uses the same structure — so Hallucination can be compared, recombined, and cited like an element on a periodic table.
- Business
- Leadership
- Government
- Healthcare
- Education
- Sales
- Marketing
- AI
- Negotiation
- Media
- Public Policy
- Relationships
- Where in our org would Hallucination most often show up unnoticed?
- Which metric, ritual, or contract clause quietly rewards Hallucination?
- If we removed every payoff for Hallucination, what behavior would replace it?
- Who benefits when Hallucination persists — and who pays the cost?
- People defend the status quo using the language of hallucination.
- Decisions cluster around the easiest narrative rather than the strongest evidence.
- New data changes the slide deck but not the decision.
- Anyone naming the pattern is treated as the problem.
Every Atlas entry is a node in a knowledge graph. See the related rail below to follow the connections.
See Hallucination through 4 lenses
Each layer of the Incentives OS reframes this concept with its own thinkers, vocabulary, and diagnostic question.
- Layer 8Organizational Psychology
What is the org actually rewarding — versus claiming to reward?
- Layer 11Economics & Mechanism Design
Who pays, who is paid, and what does the price signal hide?
- Layer 13Leadership
What kind of leadership move does this situation actually require?
- Layer 15AI & Alignment
What proxy reward is the AI optimizing — and what is it ignoring?
Do you actually know Hallucination?
Three quick questions. Result is saved into your review streak — come back when the term is due to lock it in.
Which best describes Hallucination?
Worked example, counter-example & concept map
On-demand AI analysis grounded in the Lab's research. Cached on your device after first run.
When you encounter Hallucination, your amygdala tags it as threat before your reasoning brain even knows what happened — and threat wins the first move.
Threat detection, fear, social pain, loss aversion, fast emotional tagging. Loss feels roughly twice as bad as equivalent gain feels good. Social rejection lights up the same circuits as physical pain.
See Amygdala in the Brain Atlas →Picked for you, from the Atlas
Ranked by shared learning paths, overlapping chips, and what you've saved.
Treating AI as more humanlike than it is.
Reduced vigilance with automated systems.
Quantifying uncertainty in model outputs.
Matching trust in a system to its actual reliability.
Written rules about how AI may be used internally.
Innovators → early adopters → majority → laggards, AI-specific.
Send the card, not just the link
A pre-rendered social card with the title, eyebrow, and URL. Copy the link, post it anywhere, or download the SVG for slides.
More definitions to follow
Every term in the Atlas connects to a dozen others. Pick any of these and see where it takes you.
The integrative, systemic, flexible value system — the first of second-tier.
It's right because it's how we've always done it.
Candidates dependent on large donors become structurally responsive to donor priorities over voter priorities.
Preferences aren't retrieved from a stable internal list — they're built on the spot from cues, context, and the elicitation method.
Multitasking measurably degrades performance.
Crediting a source or cause that isn't actually responsible.
One strong trait colors judgment of unrelated traits.
Believing people get what they deserve.
Anything that can go wrong will go wrong.
Imagine the project failed; explain why.
Teams under-reporting AI capability to protect comp or status.
The conditions under which a target accepts an influence attempt without scrutiny.