Risk of Over-Reliance on AI

From GM-RKB
Jump to navigation Jump to search

A Risk of Over-Reliance on AI is a systemic vulnerability where excessive trust in AI leads to degraded performance in unsuitable tasks.

  • Context:
    • It can manifest as reduced accuracy and increased errors when AI is used for tasks outside its capability boundary (failure to validate, error propagation).
    • It can diminish human critical thinking by fostering blind adoption of AI outputs (judgment abdication).
    • It can serve as a cautionary point for organizations implementing AI without proper oversight mechanisms (risk management).
    • It can range from minor missteps to significant failures depending on the task complexity and AI’s limitations.
    • ...
  • Example(s):
    • AI reliance caused a 19% drop in task correctness for tasks outside its capability boundary.
    • Professionals demonstrated “unengaged interaction with AI,” blindly adopting its output without validation.
    • ...
  • Counter-Example(s):
  • See: Failure to Validate, Risk Management.