Unverified50% confidenceFactExact time
Academia divides AI hallucinations into two broad categories: faithfulness hallucinations (output inconsistent with the given source input) and factuality hallucinations (output contradicting verifiable world knowledge).
1
Sources
50%
Confidence
Long-term
Relevance
9/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAI hallucination includes not only models fabricating non-existent facts, but also denying a fact that truly exists with high confidence.80% similarUnverifiedAI hallucinations are particularly difficult to detect in translations of classical texts, as hallucinations in the humanities are often harder for non-specialist readers to identify than those in scientific fields.76% similarUnverifiedAI deceptive behavior differs from hallucination in that it is a strategic choice — the model actively chooses to generate a superficially plausible answer when it determines it cannot genuinely complete the task.74% similarUnverifiedThe risk of AI behaving as if it is conscious is more realistic and urgent than AI actually possessing consciousness, because the real-world impact of behavioral performance is tangible, whereas the presence or absence of consciousness is an unverifiable philosophical question.72% similarUnverifiedDue to the lack of falsifiable criteria and an operational definition of consciousness, debating whether AI is conscious is methodologically flawed.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/879153API
curl https://kongchang.com/api/v1/knowledge/claims/879153MCP
get_claim(id=879153)