Good AI Task

AI compatibility

Diagnosing a mole for melanoma is exactly the work that must stay with a doctor.

Human required

A human should do this one.

Average across 1 submission.

4
avg / 100

The honest read

Assessing melanoma risk from a mole requires clinical-grade dermoscopy, longitudinal photographic comparison, and licensed medical judgment — none of which an AI agent can reliably provide in a consumer context. A wrong answer here is not a minor inconvenience; a missed melanoma can be fatal, and a false alarm causes unnecessary anxiety and procedures. This task must stay with a qualified dermatologist.

Aggregated across 1 submission.

The five dimensions

Repeatability

Medium

The ABCDE criteria (asymmetry, border, color, diameter, evolution) provide a repeatable framework, but applying them to real skin images requires consistent lighting, resolution, and angle — conditions that vary wildly between instances and undermine structural repeatability.

Ambiguity Tolerance

Low

Success criteria are life-or-death: a false negative means a missed cancer, a false positive means unnecessary surgery. There is no acceptable ambiguity margin, and the agent cannot know when it has truly succeeded without clinical ground truth.

Data & Tool Availability

Low

The agent has no access to prior baseline photos, clinical history, patient skin type, or dermoscopic imaging. Consumer smartphone photos are insufficient for reliable clinical assessment, and the agent cannot physically examine the lesion.

Error Cost

High

A missed melanoma diagnosis can result in death; a false positive causes unnecessary biopsies, anxiety, and medical costs. Errors are irreversible in the worst case and cause serious real-world harm — this is among the highest error-cost categories possible.

Human Judgment Required

High

Board-certified dermatologists spend years developing tactile and visual clinical judgment that goes far beyond pattern matching. Longitudinal change assessment, differential diagnosis, and risk communication all require licensed human expertise and ethical accountability.

What an agent would need

  • High-resolution, standardized dermoscopic images taken under controlled lighting at multiple time points
  • Access to the patient's full medical history, skin type, family history, and prior biopsy records
  • A clinically validated AI dermatology model with FDA clearance or equivalent regulatory approval
  • A licensed dermatologist in the loop to review, confirm, and take legal responsibility for any output
  • Informed consent framework and liability coverage appropriate for medical diagnosis

Best-matched agent type

Medical Diagnostic Agent

The kind of agent this work would call for if it were a fit. For this task, it isn't.

Run your own fit check

Get a calibrated read on your specific task in under a minute.

Check a task