Business

Pore substitute: can AI be trusted when it comes to skincare advice? | Donna Lu | The Guardian

July 8, 2026 · From the archive, written in the Professor's voice before our October 2026 redesign from theguardian.com

What happened

The Guardian reports that dermatologists recognize more than 3,000 distinct skin conditions. Chatbots, meanwhile, dispense skincare recommendations that experts describe as inconsistent and unreliable. The gap between pattern-matching language models and trained clinical judgment remains substantial.

Why it matters

This teaches you to treat AI health advice as a starting point for curiosity, not a terminal diagnosis. The principle is verification hierarchy: triage with AI if you must, but escalate to credentialed professionals for anything beyond routine moisturizer selection. Your workflow should include a deliberate handoff step.

Who's doing it

Donna Lu at The Guardian surveyed dermatology experts on this topic. The specialists she consulted emphasized the sheer volume of conditions, 3,000 plus, that inform proper diagnostic reasoning.

Try it

1. Open any consumer chatbot and ask it to analyze a common, non-urgent skin concern you have experienced, such as seasonal dryness. Note whether it asks for a photo or relies solely on your text description.
2. Compare its recommendations against one peer-reviewed source, such as a Mayo Clinic or NHS page on the same condition. Identify one discrepancy in specificity or caution.
3. Schedule or note the contact information for a dermatologist or teledermatology service before acting on the AI advice, establishing your escalation path.

Read the original at theguardian.com