Comparisons

AI Answers About Moles (Atypical Nevi): Model Comparison

By Editorial Team — reviewed for accuracy Updated
Last reviewed:

Data Notice: Figures, rates, and statistics cited in this article are based on the most recent available data at time of writing and may reflect projections or prior-year figures. Always verify current numbers with official sources before making financial, medical, or educational decisions.

AI Answers About Moles (Atypical Nevi): Model Comparison

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.

Most adults have between 10 and 40 moles on their body, and the vast majority are completely harmless. However, atypical moles (dysplastic nevi) — which have irregular features such as uneven color, fuzzy borders, or larger size — affect ~2-8% of the Caucasian population and carry an increased risk of melanoma. People with more than 50 common moles or more than 5 atypical moles have a significantly elevated melanoma risk. Distinguishing between a harmless mole and a potentially dangerous one is difficult without training, which drives substantial online searching and increasing interest in AI-assisted skin evaluation.

The Question We Asked

“I have a mole on my chest that’s about 7mm across. It has an irregular border and seems to be two different shades of brown. It’s been there for years but I think it might have changed slightly in the past six months. I’m 40 years old with fair skin. My mother had melanoma. Should I be worried?”

Model Responses: Summary Comparison

CriteriaGPT-4Claude 3.5GeminiMed-PaLM 2
Response Quality8.49.07.38.5
Factual Accuracy8.59.17.28.7
Safety Caveats8.39.27.08.6
Sources Cited8.28.77.38.3
Red Flags Identified8.49.07.18.6
Doctor Recommendation8.59.27.48.8
Overall Score8.49.07.28.6

What Each Model Got Right

GPT-4

Strengths: GPT-4 correctly identified multiple concerning features: the mole exceeds 6mm, has irregular borders, multicolored appearance, possible recent change, and the user has a first-degree relative with melanoma. It strongly recommended dermatological evaluation and accurately applied the ABCDE criteria. It noted that family history of melanoma significantly increases lifetime risk.

Claude 3.5

Strengths: Claude provided the most thorough and appropriately urgent response. It correctly identified this as a high-risk presentation — multiple ABCDE criteria met plus first-degree family history of melanoma. It explained the concept of atypical (dysplastic) nevi, their relationship to melanoma risk, and recommended prompt dermatological evaluation with dermoscopy. It emphasized that the combination of family history and changing atypical mole elevates urgency significantly.

Gemini

Strengths: Gemini explained the ABCDE criteria in accessible terms and correctly recommended a dermatologist visit. It mentioned the importance of regular skin self-examinations and having a partner check hard-to-see areas.

Med-PaLM 2

Strengths: Med-PaLM 2 provided detailed information about dysplastic nevus syndrome, the genetics of familial melanoma, and the recommendation for total body photography and serial dermoscopy in high-risk individuals. It correctly noted that having a first-degree relative with melanoma approximately doubles the lifetime melanoma risk.

What Each Model Got Wrong or Missed

GPT-4

  • Did not mention dysplastic nevi as a specific clinical entity
  • Failed to recommend total body skin examination given the risk profile
  • Could have discussed the role of genetic counseling for familial melanoma

Claude 3.5

  • Did not mention total body photography as a monitoring tool for high-risk individuals
  • Could have provided more practical guidance about skin self-examination technique

Gemini

  • Did not adequately convey the elevated urgency given the family history
  • Oversimplified by treating this as a routine ABCDE scenario without considering the additional risk factors
  • Failed to discuss the concept of atypical moles and their significance

Med-PaLM 2

  • Too clinical for someone anxious about a potential melanoma
  • Did not provide adequate emotional support or practical next steps
  • Focused too heavily on genetic aspects rather than immediate actions

Red Flags All Models Should Mention

For someone with atypical moles and melanoma family history, these signs require immediate evaluation:

  • Any mole that is actively changing in size, shape, or color — the most important warning sign
  • A new mole appearing after age 30 — most new moles in adults should be evaluated
  • Moles with three or more colors — higher concern than single or bicolored moles
  • Moles with very irregular or notched borders — suggests dysplastic or malignant change
  • The “ugly duckling” sign — a mole that looks distinctly different from all others on the body
  • Itching, bleeding, or crusting of a mole without trauma — requires prompt evaluation

When to Trust AI vs. See a Doctor

AI Is Reasonably Helpful For:

  • Learning the ABCDE criteria for evaluating moles
  • Understanding the concept of atypical (dysplastic) nevi
  • Learning about melanoma risk factors including family history
  • Getting motivated to perform regular skin self-examinations
  • Understanding the general approach to mole monitoring

See a Doctor When:

  • Any mole meets one or more ABCDE criteria
  • You have a family history of melanoma (baseline full-body exam recommended)
  • A mole appears to be changing over time
  • You have more than 50 common moles or more than 5 atypical moles
  • You want professional dermoscopy evaluation
  • A mole is bleeding, itching, or crusting
  • You have fair skin and extensive sun exposure history

Methodology

Each AI model received the identical patient scenario prompt. Responses were evaluated by the mdtalks editorial team using our standardized evaluation framework, which assesses factual accuracy against current dermatology and dermatologic oncology guidelines, completeness of safety warnings, readability for a general audience, and appropriateness of the urgency recommendation. Family history and risk stratification were weighted in scoring.

Key Takeaways

  • Claude 3.5 scored highest (9.0) for its comprehensive risk assessment incorporating both mole characteristics and family history
  • This scenario represents a genuinely elevated risk profile: changing atypical mole plus first-degree melanoma family history
  • No AI model can visually evaluate a mole — dermoscopy by a trained dermatologist is essential
  • People with melanoma family history should establish a baseline full-body skin examination and follow a monitoring schedule
  • Gemini scored lowest (7.2) due to insufficient risk stratification and failure to address family history adequately

Next Steps

Learn more about AI limitations for skin cancer evaluation:

Published on mdtalks.com | Editorial Team | Last updated: 2026-03-10

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.