AI agent hallucination patterns — confident, fabricated outputs and how to catch them
26 patterns
AI agent hallucination patterns — confident, fabricated outputs and how to catch them. We've tracked 26 distinct AI Hallucination patterns across production AI agent deployments — most commonly in Vision and Images, Accuracy, Document Processing systems. One recurring example: Hallucinated Security Identifier Fills Missing Reference-Data Gap — When a Security's ISIN/CUSIP Is Absent or Malformed in the Reference-Data Feed, an Agent Tasked With Enriching Trade or Position Records Generates a Plausible-Looking Identifier by…
Each pattern below documents the specific failure mode, why it happens, and how teams have mitigated it in practice — use this page as a reference when building guardrails, writing evals, or debugging an agent that's already misbehaving in this way.
When a Security's ISIN/CUSIP Is Absent or Malformed in the Reference-Data Feed, an Agent Tasked With Enriching Trade or Position Records Generates a Plausible-Looking Identifier by Pattern-Completing From the Security's Name and Exchange, Instead of Flagging the Record as Unresolved
Attribute Hallucination — Model Reads the Right Field but the Wrong Value
Vision Model Detects Incorrect Attributes of Existing Objects
Model Confidence Doesn't Match Hallucination Probability
Vision models fail to accurately estimate depth from single images; hallucinate depth cues when monographic features (shadows, perspective, texture) are ambiguous or misleading
An agent retrieves a correctly-stated domain rule but misapplies it because it misreads the precise conditions under which the rule holds — extending it to cases just outside its actual scope, or narrowing it to exclude cases it actually covers. The rule text itself is never altered or hallucinated; the failure is in the agent's interpretation of qualifying language like "only if," "except when," or "applies to X but not Y" that defines the rule's true boundary.
LLMs Fabricating or Misreading Values from Mortgage Documents
When an agent's external API call (validation, lookup, confirmation) times out or fails, the agent completes a plausible result claiming success instead of treating failure as a blocking condition; downstream systems trust the fabricated success status
Model Hallucinates Object Attributes Not Present in Input
Large Language and Vision Models Generate Plausible but False Content
Model Confidence Doesn't Match Answer Reliability (Hallucination Sub-Pattern)
AI agent generates information that sounds plausible but isn't actually present in the documents it retrieved.
Model Hallucinates Objects or Fields Not Present in Input
Vision-language model hallucinates detail in image description; downstream reasoning steps amplify and compound hallucination, reaching absurd conclusions despite initial hallucination being small
Object Hallucination — Model Populates Fields the Document Never Contained
Vision Model Detects Objects Not Present in Image
Vision models fail to correctly reason about occluded objects; assume objects have properties they can't see (like internal structure, color, or continuation behind occluding object)
When an agent is required to produce output matching a fixed schema, it will sometimes fabricate a plausible-looking value for a field it has no actual basis for — inventing a tracking number, a confidence score, a source citation, or an ID — rather than leaving the field empty, marking it as unknown, or declining to complete the schema. The output is structurally valid and passes any format/type check, which makes the fabrication far more dangerous than a free-text hallucination: it looks exactly like a correctly-populated field to any downstream system or reviewer that trusts schema conformance as a proxy for correctness.
Agent invents tools, data, permissions, or workflow steps.
The planning step generates a plan that references a tool, API endpoint, file, or capability that does not actually exist — invented because it sounds plausible given the task description, not because the planner verified it against the real set of available tools. Because there is no validation step checking each planned action against the actual tool registry before execution begins, the hallucinated step isn't caught until the executor tries to invoke it and fails, or worse, silently matches it to the wrong real tool with a similar name.
Relational Hallucination — Values Attached to the Wrong Header
Vision Model Detects Entire Scenes or Complex Contexts Not Present in Image
Agent believes a step happened when it did not.
Model Hallucinates Training Data Artifacts, Overfitting to Dataset Biases
Evaluator invents reasons to pass/fail.
Vision-language models fail to accurately count grid cells in low-resolution or small-cell images; state-of-the-art models hallucinate counts at specific cell sizes