A popular skin cancer detection app, SkinVision, is facing scrutiny following a new study from Ghent University and Ghent University Hospital in Belgium. The research, published in late February 2026, indicates the app misses a significant proportion of actual skin cancers – up to one in four – and incorrectly flags benign moles as potentially cancerous in nearly one in eight cases. This raises concerns about the reliability of AI-powered diagnostic tools, particularly those offered directly to consumers and sometimes reimbursed by health insurers.
The app functions by allowing users to photograph skin lesions with their smartphones, after which an artificial intelligence algorithm analyzes the image and provides an assessment of risk. SkinVision is used by over 3 million people worldwide, with 100,000 checks completed in Belgium alone during 2025. While the company emphasizes the app is intended as a support tool and not a definitive diagnostic replacement, the study’s findings highlight potential pitfalls of relying solely on such technology for early cancer detection. The core issue surrounding skin cancer detection apps is their accuracy in real-world conditions.
Study Details and Findings
The research, conducted by UGent and UZ Gent, involved a prospective study evaluating the app’s performance with actual users and their skin lesions. Researchers found that SkinVision produced a false alarm – incorrectly identifying a benign mole as potentially cancerous – in approximately 12.5% of cases. More concerningly, the app failed to detect skin cancer in roughly 25% of instances. These findings are detailed in a report by VRT NWS, a Belgian news outlet.
The study is part of a larger doctoral research project at Ghent University, titled “AI-based tools for skin cancer detection: are we ready for them?” which began on April 28, 2025, and is scheduled to conclude on September 20, 2026. The research aims to evaluate the accuracy and safety of AI-based tools in practical settings and identify how they can best contribute to public health. A key focus is on assessing the performance of these tools at the consumer level, where accuracy can be particularly challenging to maintain. The project is being led by doctoral researcher Julie Kips.
SkinVision’s Response and Ongoing Development
SkinVision has responded to the study’s findings by asserting that the research is outdated and that the accuracy of their AI algorithm has improved year-over-year. The company maintains that the app is designed to be a supportive tool, encouraging users to consult with a healthcare professional for any suspicious lesions. According to Tweakers, the app provider claims the Belgian research is outdated.
The company also points out that the app is not a substitute for professional medical diagnosis. If the app flags a mole as potentially concerning, users are advised to promptly schedule an appointment with their doctor or a dermatologist. Conversely, if the app deems a mole benign, it can be monitored but should still be checked by a medical professional if any changes occur.
Broader Implications for AI in Healthcare
This study adds to a growing body of research examining the efficacy of AI-powered diagnostic tools in healthcare. While AI holds immense promise for improving access to care and enhancing diagnostic accuracy, concerns remain about the potential for errors and the need for rigorous validation in real-world settings. The Research Explorer project details concerns about the accuracy of many available apps. (Research Explorer – DOCT/013463)
The incident underscores the importance of critical evaluation when using such tools and the necessity of maintaining a strong doctor-patient relationship. Consumers should be aware of the limitations of AI-driven assessments and prioritize professional medical advice for any health concerns. The increasing incidence of skin cancer in Europe – and globally – further emphasizes the need for accessible and accurate detection methods, but also highlights the risks of relying on unproven technologies.
The Ghent University research team plans to continue its investigation into the performance of AI-based skin cancer detection tools, with a focus on identifying factors that influence accuracy and user acceptance. The next phase of the study will explore how these tools can be integrated into existing healthcare systems to optimize their impact on public health. The project is expected to publish further findings in September 2026.
Have your own experience with skin cancer detection apps? Share your thoughts in the comments below, and please share this article with anyone who might locate it helpful.
Disclaimer: This article provides information for general knowledge and informational purposes only, and does not constitute medical advice. It is essential to consult with a qualified healthcare professional for any health concerns or before making any decisions related to your health or treatment.
