Which risk category addresses situations where AI systems behave in ways inconsistent with their intended objectives?

Question

Options

• AData leakage risk
• BMisalignment risk
• CAvailability risk
• DVendor lock-in risk

• A
  5% (1)
• B
  86% (18)
• C
  5% (1)
• D
  5% (1)

Explanation

Misalignment risk (B) is correct because it specifically describes the failure mode where an AI system's actual behavior diverges from its designers' intended goals - for example, an AI optimizing for a proxy metric in ways that undermine the true objective.

• A (Data leakage risk) is wrong - that refers to sensitive data being exposed or exfiltrated, not behavioral divergence from goals.
• C (Availability risk) is wrong - that concerns system downtime or inaccessibility, not goal misalignment.
• D (Vendor lock-in risk) is wrong - that's a business/procurement concern about dependency on a single provider, unrelated to AI behavior.

Memory tip: Think of "misalignment" like a car with a misaligned wheel - it's still running, but heading somewhere other than where you intended. The "mis-" prefix (meaning "wrongly") applied to "alignment" (goals matching behavior) makes B self-defining.

No community discussion yet for this question.