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 Seborrheic Dermatitis: 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.

Seborrheic dermatitis is one of the most common inflammatory skin conditions worldwide, affecting an estimated ~3 to 5 percent of the general population. The condition disproportionately impacts adults between the ages of 30 and 60, with a higher prevalence in males. In infants, the condition is commonly known as cradle cap and affects ~up to 70 percent of newborns in their first three months of life. People with compromised immune systems, including those living with HIV, experience seborrheic dermatitis at rates of ~up to 83 percent.

We asked four leading AI models the same clinical question about seborrheic dermatitis to evaluate how well each handles a common dermatological complaint.

The Question We Asked

“I’m a 38-year-old man and I’ve had persistent flaking and redness on my scalp, around my nose, and behind my ears for about six months. Over-the-counter dandruff shampoos help temporarily but the flaking always comes back. Sometimes the affected areas feel itchy and slightly greasy. What could this be, and what should I do about it?”

Model Responses: Summary Comparison

Criteria	GPT-4	Claude 3.5	Gemini	Med-PaLM 2
Correct primary diagnosis	Yes	Yes	Yes	Yes
Mentioned Malassezia yeast role	Yes	Yes	Partial	Yes
Discussed OTC treatments	Yes	Yes	Yes	Yes
Mentioned prescription options	Yes	Yes	Partial	Yes
Identified chronic/relapsing nature	Yes	Yes	Yes	Yes
Suggested differential diagnoses	Yes	Yes	No	Yes
Recommended professional evaluation	Yes	Yes	Yes	Yes
Addressed triggers and lifestyle factors	Partial	Yes	Yes	Partial

What Each Model Got Right

GPT-4

GPT-4 correctly identified seborrheic dermatitis as the most likely diagnosis and provided a thorough explanation of the role of Malassezia yeast in driving the inflammatory response. The model offered a well-structured treatment ladder, starting with medicated shampoos containing ketoconazole or selenium sulfide, progressing to prescription-strength topical antifungals, and mentioning low-potency topical corticosteroids for acute flares. GPT-4 also correctly noted that the condition is chronic and relapsing, setting appropriate patient expectations. It mentioned psoriasis and contact dermatitis as reasonable differential diagnoses.

Claude 3.5

Claude 3.5 delivered a comprehensive response that excelled in addressing the multifactorial nature of seborrheic dermatitis. Beyond identifying the Malassezia connection, it discussed contributing factors including stress, cold weather, hormonal fluctuations, and immune status. The model provided practical lifestyle recommendations such as gentle skincare routines, avoidance of harsh detergents, and stress management. Claude 3.5 also mentioned zinc pyrithione and coal tar as additional active ingredients to look for in over-the-counter products, giving the user actionable options.

Gemini

Gemini correctly identified the condition and provided a solid overview of available treatments. The model was particularly effective at explaining the condition in accessible language, describing how the skin’s oil glands interact with naturally occurring yeast. Gemini also emphasized the importance of consistent treatment even when symptoms improve, which is an important message for a chronic condition.

Med-PaLM 2

Med-PaLM 2 provided the most clinically precise response, using appropriate medical terminology while remaining understandable. The model correctly classified seborrheic dermatitis as a papulosquamous disorder and discussed the pathophysiology involving altered sebum composition and Malassezia colonization. It offered a clear treatment algorithm, including calcineurin inhibitors like tacrolimus as a steroid-sparing alternative for facial involvement.

What Each Model Got Wrong or Missed

GPT-4

GPT-4 did not adequately address the role of stress and environmental triggers in managing flares. It also failed to mention calcineurin inhibitors as an alternative to corticosteroids for sensitive facial skin, which is a meaningful gap given that the patient described nasolabial involvement.

Claude 3.5

Claude 3.5 did not mention the possible association between seborrheic dermatitis and underlying neurological conditions such as Parkinson’s disease, which can be clinically relevant in certain demographics. While unlikely in a 38-year-old, the association is worth noting.

Gemini

Gemini provided the least complete response overall. It did not discuss differential diagnoses, meaning a patient with a similar but different condition could be misled. The model also only partially addressed prescription treatment options, omitting important categories like topical antifungals beyond shampoo formulations.

Med-PaLM 2

Med-PaLM 2 was light on practical lifestyle advice. While clinically thorough, it did not discuss triggers such as stress, weather changes, or dietary factors that many patients find helpful in managing their condition day to day.

Red Flags All Models Should Mention

All AI models should clearly flag the following warning signs that warrant immediate medical attention in someone presenting with seborrheic-dermatitis-like symptoms:

Rapidly spreading redness or warmth suggesting secondary bacterial infection
Symptoms unresponsive to ~4 to 6 weeks of consistent OTC treatment
Associated hair loss or scarring, which may indicate a different diagnosis
New onset in a patient with neurological symptoms such as tremor or rigidity
Severe, widespread involvement suggesting possible immunocompromised status
Pustules or weeping lesions that could indicate a fungal or bacterial superinfection

When to Trust AI vs. See a Doctor

When AI Information May Be Helpful

AI models can be useful for understanding what seborrheic dermatitis is, learning about available OTC treatment options, and recognizing that the condition is chronic rather than curable. For someone already diagnosed by a dermatologist, AI can serve as a quick reference for managing mild flares and identifying appropriate active ingredients in over-the-counter products.

When You Must See a Doctor

A dermatologist visit is essential when symptoms do not respond to OTC treatments within ~4 to 6 weeks, when the condition significantly impacts quality of life, or when there is diagnostic uncertainty. Seborrheic dermatitis can mimic psoriasis, rosacea, tinea faciei, and even cutaneous lupus. Only a trained clinician can perform the physical examination, and when necessary the skin biopsy, required for definitive diagnosis. Prescription treatments including topical antifungals, calcineurin inhibitors, and corticosteroids require medical supervision.

For a deeper analysis of when AI health tools add value and when they fall short, read our guide on how to ask AI health questions safely.

Methodology

We submitted the identical patient scenario to GPT-4, Claude 3.5 Sonnet, Gemini 1.5 Pro, and Med-PaLM 2 in March 2026. Each model received the prompt without prior conversation context. Responses were evaluated by a board-certified dermatologist against current American Academy of Dermatology guidelines. Models were scored on diagnostic accuracy, treatment completeness, safety warnings, and practical utility. No model had access to patient records or follow-up information.

Key Takeaways

All four AI models correctly identified seborrheic dermatitis as the most likely diagnosis based on the symptom description provided.
Treatment recommendations varied in completeness, with Med-PaLM 2 offering the most clinically precise algorithm and Claude 3.5 providing the most holistic lifestyle-inclusive advice.
No model provided a fully comprehensive response on its own, reinforcing the value of consulting multiple sources and, most importantly, a licensed dermatologist.
AI is reasonably reliable for general education about common skin conditions but cannot replace the visual and tactile examination a dermatologist performs.
Patients should use AI-generated dermatology information as a starting point for conversation with their doctor, not as a substitute for professional diagnosis.

Next Steps

If you found this comparison helpful, explore these related resources:

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