SAVE 30% ON EACH EBOOK AND 60% ON THE COMPLETE BUNDLE!
Can AI Diagnose HSV From a Photo? What AiSTI Research Shows
AiSTI is a research-stage lesion-screening tool, not a consumer HSV test. Learn what studies show, the limits of photo analysis and when to seek a swab.
TECH & DIGITAL HEALTH
Brandon
1/30/20265 min read
A photo cannot confirm whether a lesion is herpes. AiSTI is a research-stage tool that analyses anogenital lesion images to screen for possible sexually transmitted infections. It may eventually help direct people to timely care; a new or unexplained lesion still needs clinical assessment and, where appropriate, a swab.
This guide explains what researchers have actually studied, what the results mean, and what to do if you have a new or unexplained lesion.
What AiSTI actually is
AiSTI is being developed by researchers at Melbourne Sexual Health Centre and Monash University in Australia. It is a web-based image-recognition system that combines a photograph of a visible anogenital lesion with information supplied by the user.
The aim is to distinguish lesions that may be related to an STI from those that are more likely to have another cause, then support triage and appropriate care. It was not designed as an HSV-only laboratory test. The official project page describes a public-facing version as being developed; it does not establish that a clinically validated consumer service is currently available.
What the main AiSTI study found
The main published evaluation used 4,913 clinical images: 1,583 images of STI-related lesions and 3,330 images of non-STI lesions. Among the STI images, the diagnoses included syphilis, genital warts and herpes; herpes accounted for 19.4% of that group.
The image-only model achieved an area under the receiver operating characteristic curve, or AUC, of 0.859. When clinical metadata was added, the AUC rose to 0.893.
An AUC of 0.893 does not mean the tool is “89.3% accurate.” AUC measures how well a model separates two groups across many possible thresholds. It is not the same as accuracy, sensitivity or specificity at a chosen clinical cut-off. The study evaluated classification of STI-related versus non-STI lesions; it did not establish 90% sensitivity and 94% specificity for diagnosing HSV.
The evaluation was retrospective and based on data from one clinical service. The authors said that further development and evaluation using larger datasets were needed. External validation across different populations, devices, lighting conditions and skin tones is especially important before a tool like this could be relied on in routine care.
What Later Studies Examined
Later AiSTI research asked 12 people about using a prototype and tested how 512 participants said they might respond to example result screens. These studies help researchers design clearer triage messages. They did not measure diagnostic accuracy in everyday users or show that the tool improves health outcomes.
A separate study presented hypothetical result screens to 512 participants and measured intended care-seeking behaviour. It examined how people might respond to different risk messages. It did not test whether AiSTI accurately diagnosed real users in unsupervised, everyday conditions.
Together, these studies show active research into usability and triage. They do not establish real-world consumer safety, diagnostic performance or improved health outcomes.
Claims the evidence does not support
The published papers and official project information reviewed for this update do not substantiate claims that AiSTI:
confirms HSV from a photograph in 30 seconds;
has 90% sensitivity and 94% specificity for HSV;
is more accurate than an in-person clinician;
is currently sold to consumers for a stated per-scan or subscription price;
is available through SH:24, Bupa or other named consumer partners;
automatically deletes images within a fixed period or guarantees anonymity; or
is safe to use as the basis for starting treatment or making disclosure decisions.
Those details should not be assumed unless the operator publishes current validation, regulatory, availability and privacy information that directly supports them.
Why a photo cannot diagnose HSV
Herpes can resemble folliculitis, ingrown hairs, contact dermatitis, fungal irritation, syphilis and other conditions. Even an experienced clinician may find visual diagnosis difficult. A photograph also cannot capture the full history, pain, tingling, fever, lesion evolution or other examination findings.
A photo cannot detect asymptomatic infection, and appearance alone cannot reliably confirm whether a lesion is caused by HSV-1 or HSV-2. Healing or older lesions can be even harder to assess.
The US Centers for Disease Control and Prevention recommends confirming suspected genital herpes from an active lesion using a type-specific nucleic acid amplification test, usually PCR, or viral culture. PCR is generally the most sensitive lesion test. A negative result from an older or healing lesion does not always exclude infection because viral shedding is intermittent.
What to do if you have a new lesion
Seek clinical care promptly. A fresh blister or ulcer offers the best chance of obtaining a useful swab.
Ask whether a PCR or NAAT swab is appropriate. This can confirm HSV and identify the type when viral material is detected.
Avoid sexual or close skin-to-skin contact involving the affected area until you have been assessed and the lesion has healed.
Do not start, stop or change antiviral treatment solely because of a photo app. Discuss treatment with a qualified clinician.
Use blood tests carefully. They cannot confirm that a particular lesion is herpes and have important timing and false-positive limitations.
For more detail, read our guides to genital herpes symptoms and outbreak timing, what at-home HSV tests can and cannot tell you, and valacyclovir and acyclovir.
AI may still help—but as triage
For example, a future tool might flag that a new sore needs a clinic swab and help a user find a service. It could also support clinicians where specialist advice is scarce. Before routine use, developers need prospective testing in real settings, external validation across skin tones and devices, clear handling of uncertain results, privacy protections and a route to confirmatory care. Research on AI assistance with HSV Western blot interpretation is a separate laboratory use and should not be confused with lesion-photo screening.
Before routine use, developers need prospective studies in real-world settings, external validation, clear performance at clinically chosen thresholds, representative data across skin tones, transparent handling of uncertain results, privacy protections and a safe pathway to confirmatory testing and treatment.
AI is also being studied elsewhere in HSV care. For example, researchers have evaluated AI assistance for interpreting HSV Western blot results in laboratories. That is a different use case from diagnosing a consumer’s lesion photograph and should not be confused with AiSTI.
Privacy questions before uploading intimate photos
An intimate image is sensitive health data. Before using any service, look for clear answers to these questions:
Who operates the service and in which country?
Where are images stored and encrypted?
Can staff, contractors or clinicians view them?
How long are images and account data retained?
Can you request deletion, and what exceptions apply?
Will images be reused to train or test an algorithm?
Which privacy law and complaints process apply?
Do not assume that an app is anonymous, clinically regulated or confidential simply because it describes itself as AI-powered.
Bottom line
AiSTI is credible research into image-based STI screening, led by Melbourne Sexual Health Centre and Monash University. The results are promising for distinguishing possible STI-related lesions from other lesions, especially when an image is combined with clinical information.
It is not evidence that a phone photo can diagnose HSV, replace a clinician or replace a lesion swab. If you have a fresh blister, ulcer or unexplained genital lesion, seek prompt clinical assessment and ask about PCR testing.
You can find a broader evidence-based overview in The Modern Man’s Guide to HSV.
Sources
Melbourne Sexual Health Centre: Image Capture Study and AiSTI project
BMC Medicine: Evaluation of AI-powered screening for STI-related skin lesions
Qualitative study of user perspectives on the AiSTI prototype
Vignette study of intended care-seeking after AI-generated results
Medical review note: This article is educational and does not provide a diagnosis. Evidence and official project information were checked on 21 September 2026.
MODERN HSV PLAYBOOK
THE SERIES
© 2026 Modern HSV Playbook. All rights reserved.
SITE
Content for informational purposes only. Not a substitute for professional medical advice.