Press & Media Kit
ScanSkinAI Press & Media Kit
Everything you need to write about ScanSkinAI accurately: approved descriptions, verifiable facts, statistics with their sources, our clinical reviewers, and logo files you can use without asking.
- Audited accuracy
- 95.3%
- Conditions covered
- 80
- Clinical reviewers
- 11
- Response time
- 1 working day
Approved descriptions
Use either of these verbatim. Both are kept current, so a copy taken from here will not go stale mid-story.
One line
ScanSkinAI is an AI skin-screening platform that lets anyone photograph a mole, rash or skin concern and get a preliminary risk assessment in seconds, with optional review by a licensed dermatologist.
One paragraph
ScanSkinAI, built by Ivy AI Solutions Limited, is an AI-assisted skin-screening platform used to check moles, rashes and other skin concerns from a photograph. The free AI check returns a preliminary risk assessment and a recommended next step; users who want a clinical opinion can request a written review from a licensed dermatologist, and users tracking a changing lesion can monitor it photographically over three months. ScanSkinAI is a UKCA Class I medical device and publishes its validation results, model methodology and aggregate usage data as citable, DOI-archived reports. It provides preliminary screening information and is not a medical diagnosis.
Company facts
- Product name
- ScanSkinAI
- Legal entity
- Ivy AI Solutions Limited
- Founder
- Dr. Lifeng Zhu, PhD
- Category
- AI-assisted skin screening and photographic mole monitoring
- Regulatory status
- UKCA Class I medical device
- Quality frameworks
- ISO 13485 quality management, ISO 27001 information security
- Model architecture
- DINOv2-Large vision transformer (ViT-L/14) at 518×518 resolution
- Availability
- Web app, installable as a PWA on iOS and Android — no app-store download required
- Media contact
- info@scanskinai.com
Statistics you can quote
Each figure names the published report it comes from. Please link to that report when you use it.
95.3%
Clinically acceptable accuracy
In an independent 100-case dermatologist audit covering 80 skin conditions, 95.3% of ScanSkinAI outputs were judged clinically acceptable — meaning the output would not lead to patient harm or a materially incorrect clinical pathway.
Source: ScanSkinAI Accuracy Report 2026
2%
Under-triage rate
Under-triage — recommending a lower urgency than appropriate, the only triage error that can delay care — occurred in 2 of 100 audited cases. Over-triage, the safe direction of error, occurred in 8.
Source: ScanSkinAI Accuracy Report 2026
I–VI
Fitzpatrick skin types validated
The audit was stratified across all six Fitzpatrick skin types and deliberately weighted toward Types III–IV, rather than validating on lighter skin only.
Source: ScanSkinAI Accuracy Report 2026
50,265
Scan records analysed
The Digital Skin Health Insights Report 2026 is a retrospective descriptive analysis of 50,265 historical scan-level risk-band records.
45,270
Medical skin-module AI outputs analysed
Alongside the risk-band records, the report analyses 45,270 medical skin-module AI-output records, covering roughly 430 distinct output labels.
2
Permanent Zenodo DOIs
The Cancer Flag Module technical whitepaper (10.5281/zenodo.21225287) and the Digital Skin Health Insights Report 2026 (10.5281/zenodo.21669080) are archived with permanent DOIs and are free to cite.
Story angles we can support
Each angle below is backed by something we have already published, so we can supply the underlying figures, method notes or a named commentator on request.
Can an app tell if a mole is dangerous?
The consumer question behind most skin-check coverage. We can speak to what AI screening can and cannot do, why 'screening' and 'diagnosis' are different claims, and how an audited accuracy figure should be read.
- 95.3% clinically acceptable outputs across 80 conditions, not a single-disease benchmark
- Why over-triage (8%) is a safer failure mode than under-triage (2%)
- What a UKCA Class I classification does and does not certify
Skin-tone bias in dermatology AI
Most published dermatology datasets skew toward lighter skin. We validated across all six Fitzpatrick types and deliberately weighted the audit toward Types III–IV, so we can talk concretely about how stratified validation is done.
- Case counts per Fitzpatrick type in the audit (I: 7, II: 14, III: 21, IV: 20, V: 13, VI: 5)
- Why aggregate accuracy can hide per-group failure
- What patients with darker skin should look for when a lesion changes
UV exposure, tanning and burn time
Seasonal peg for summer and winter-sun coverage. We publish typical clear-sky peak UV values for 48+ countries with time-to-redness estimates for the fastest-burning skin types.
- Country-by-country peak UV comparisons, embeddable as a free chart
- Why 'burn time' figures are a conservative floor, not an average
- Winter-sun destinations with UV levels most travellers underestimate
Waiting lists and the self-check gap
How people behave while waiting for a dermatology appointment, and where a photographic record changes the consultation. Useful for health-system and NHS-pressure stories.
- Photographic monitoring over three months as pre-appointment evidence
- What a clinician actually needs to see in a patient's phone photos
- Self-selection: platform data describes who worries, not population prevalence
Spokespeople and interviews
Dr. Lifeng Zhu, PhD
Founder, Ivy AI Solutions Limited
Available to discuss model architecture, validation methodology, skin-tone stratified evaluation and how AI screening fits alongside clinical dermatology.
Clinical commentary
Dermatologists from the review roster
For patient-facing questions — what a changing mole looks like, when to seek an in-person check, how self-examination should be done — we can route your question to a licensed reviewer.
Email info@scanskinai.com with your outlet, angle and deadline. Put a same-day deadline in the subject line and we will prioritise it.
Clinical reviewers
Dermatologist reviews on ScanSkinAI are written by licensed clinicians. Full biographies, credentials and licence numbers are published on the dermatologist review page.
Dr. Jessica R. Burgy
MD, Board-Certified Dermatologist · USA
Dr. Anna H. Chacon
MD, FAAD, Board-Certified Dermatologist · USA
Dr. Alexandria Marrs
MD · USA
Dr. Paul Cellura
MD · USA
Dr. Tomasz Grzelewski
MD, PhD · UK / Poland
Dr. Celina Kazumi Iwasa
GMC-Registered · UK
Dr. Hina Mazhar
MD · UK
Dr. Anand S. Urhekar
MD Dermatology · India
Dr. Abraham Yam
MD · Asia-Pacific
Dr. Katrina April M. David
MD · Philippines
Dr. Adriana Nicoletti
MD · Brazil / Australia
Company timeline
March 2026
Independent 100-case dermatologist audit completed
Board-certified dermatologists reviewed 100 anonymised real-world cases across 80 conditions and all six Fitzpatrick skin types.
Read the report2026
Cancer Flag Module whitepaper archived on Zenodo
Technical whitepaper published with a permanent DOI (10.5281/zenodo.21225287), free to cite.
Read the report2026
Digital Skin Health Insights Report 2026 published
Retrospective descriptive analysis of 50,265 scan-level records and 45,270 medical skin-module outputs, DOI 10.5281/zenodo.21669080.
Read the report2026
UKCA Class I medical device registration
ScanSkinAI operates under ISO 13485 quality management and ISO 27001 information security frameworks.
Logos and images
Media FAQ
Can I use ScanSkinAI's logo and statistics without asking permission?
Yes. Logos, screenshots and published statistics may be reproduced in editorial coverage of ScanSkinAI without prior permission, provided the logo is not altered and each statistic links to the report it came from.
How should ScanSkinAI be described in an article?
As an AI-assisted skin-screening platform that returns a preliminary risk assessment from a photograph, with optional written review by a licensed dermatologist. It is a UKCA Class I medical device and provides screening information — it does not diagnose skin cancer or replace a clinical examination.
What does the 95.3% accuracy figure actually measure?
It is the proportion of AI outputs judged clinically acceptable in an independent 100-case dermatologist audit across 80 skin conditions, meaning the output would not lead to patient harm or a materially incorrect clinical pathway. It describes screening performance, not diagnostic accuracy against histopathology.
Is a spokesperson available for interview?
Yes. Founder Dr. Lifeng Zhu, PhD can comment on model design, validation methodology and skin-tone stratified evaluation. Clinical questions can be routed to a dermatologist on the review roster. Email info@scanskinai.com with your deadline.
How quickly do you respond to media enquiries?
We aim to respond within one working day. If you are on a same-day deadline, put the deadline time in the subject line of your email to info@scanskinai.com.
Usage terms
- You may reproduce the logos and screenshots above in editorial coverage of ScanSkinAI without asking permission.
- Do not alter the logo's colours, proportions or lock-up, and do not present it in a way that implies ScanSkinAI endorses another product.
- When citing a statistic, please link to the report it came from — each figure on this page names its source.
- Describe ScanSkinAI as a preliminary screening tool, not a diagnostic one. It does not diagnose skin cancer or replace a clinician.
ScanSkinAI provides preliminary screening information and is not a medical diagnosis. Please avoid describing it as diagnosing skin cancer or replacing a clinical examination.
Media kit last updated 2026-08-08.