AI Disclaimer

DermaPrime uses artificial intelligence to analyse skin imagery. This page explains, in plain language, what those models do, their limits, and the safeguards we apply.

1. What the AI does

• Classification of dermoscopic and clinical images across common skin conditions.
• Longitudinal change detection across follow-up images of the same lesion.
• Non-invasive depth estimation for suspected melanoma.
• Explainability overlays highlighting regions that influenced the model's output.

2. What the AI does not do

• It does not make a medical diagnosis.
• It does not prescribe or recommend treatment.
• It does not replace a clinical examination, biopsy, or histopathology.

3. Known limitations

• Accuracy depends on image quality, lighting, focus, and framing.
• Performance can vary across skin tones, ages, and rare conditions; we run periodic bias audits and publish exclusion criteria.
• Models are probabilistic — false positives and false negatives are possible.

4. Training data and provenance

Our models are trained on a curated mix of public dermatology archives and proprietary clinical datasets. We disclose provenance, bias audits, and exclusion criteria on our Research page. Your data is used to improve our models only if you give a separate, optional consent.

5. Human oversight

AI outputs are informational. For any meaningful decision, a qualified clinician should be involved. The optional clinician-review feature lets a licensed dermatologist examine your submission and provide a written opinion.

6. Acceptance

By using analysis features you confirm that you understand these limitations and that the AI output is not a substitute for medical care.