AI flaw: Made-up words, probability illusion
AI Hallucinations refers to generative artificial intelligence
When the model generates text or answers questions, while appearing to be
logical and grammatically correct, its output may contain completely fabricated,
inaccurate, or factually incorrect information.
Causes
AI hallucinations are often caused by the model lacking relevant information
In some cases, content is generated through probabilistic selection rather than
based on real-world knowledge bases or logical reasoning, which makes its
output not only unreliable but also potentially misleading to users.
In addition to the key flaw of AI hallucination, potential shortcomings and limitations
include explainability, computational costs, data bias, real-time updates, data
security, personal privacy, malicious output, etc.
