A review of 105 acne-AI studies finds diagnosis accuracy near 87-90%, but very little testing by skin color. For teledermatology patients with brown or Black skin, that means a virtual acne assessment may be less reliable. Teledermatology means skin care through photos and video visits rather than in-person exams. The review, published in the International Journal of Dermatology in November 2025, found 96.2% of articles focused on diagnosis, yet only 13% reported patient skin color by ethnicity and 3.8% included that data, according to the scoping review.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- How accurate is acne AI right now?
- Why does performance drop on darker skin?
- Does AI help during a virtual visit?
- How can you get a clearer acne assessment?
How accurate is acne AI right now?
Ensemble models averaged 89.7% accuracy, deep learning 88.5%, and large language models 87.5% in the same review. Those scores look strong for sorting acne from other concerns and grading severity. The problem is who was counted.
When studies do not record darker tones, high averages can hide weaker checks on brown and Black skin. A 2025 systematic review of teledermatology and AI across skin tones found AI tools performed up to 15% worse on darker skin. That review also found no teledermatology study that separated diagnostic accuracy by skin type. Patients should treat an AI label as a starting point, not a final answer.
Why does performance drop on darker skin?
Acne can look different across tones, with more visible dark spots, uneven tone, and keloid-type scarring. Tools trained mainly on light-skin images have fewer examples for those patterns. Dermatologist Jules Lipoff warned in 2026 that darker tones remain underrepresented in training data.
Authors on the first FDA-authorized primary-care skin-cancer AI device made the same point, saying poor darker-skin representation worsens results and risks bias against historically underserved patients. Access adds another layer. A global dermatology-access study recommended wider frontline training plus AI and telemedicine to bring diverse-skin expertise closer to underserved patients. The aim is wider reach, but reach without balanced testing leaves a gap.
Does AI help during a virtual visit?
A store-and-forward simulation reported board-certified dermatologists had 38% top-3 accuracy and primary-care physicians 19% without AI, with both groups scoring lower on dark-skinned images, as reported by Northwestern University in February 2024, in the university summary. Adding deep-learning support raised accuracy by 33% for dermatologists and 69% for primary-care physicians, but did not remove the dark- versus light-skin gap, as reported by MDedge on the multicenter Nature Medicine study, in the MDedge report.
That result matters for acne care. Support can lift a weak photo visit, yet the lift is smaller when the tool and the clinician have less experience with darker skin. Human review remains important, especially for new, painful, deeply pigmented, or scarring breakouts.
How can you get a clearer acne assessment?
Send sharp, true-color photos. Take front, left, and right views in daylight, keep flash off, pull hair back, remove makeup and filters, and add one close-up that shows texture.
If the advice does not match your skin or symptoms, say so and request review by a dermatologist with experience in darker skin tones. Bring the same photo set to each follow-up so change is easy to track.
- Include your skin history: current products, prescription acne drugs, recent changes, and prior dark marks or scars.
- State your concern plainly: active bumps, lingering spots, pain, itch, hair loss, or scar change.
- Ask what else it could be: acne, eczema, folliculitis, rosacea, or drug reaction.
- Ask about follow-up: what photo or in-person check would confirm the plan, and what warning signs should prompt faster care.
You Might Also Like
- Skin-Tone Equity Gaps in 2026 Acne AI Research: What Developers Still Need to Prove for People Using Acne Apps
- Pediatric Acne and Rosacea at the 2026 AAD Annual Meeting: What the Update Means for Pediatric Dermatology Patients
- Recovery and Pigment Safety in the 2026 Split-Face Acne Scar Trial: What the Trial Means for People With Rolling Acne Scars