{"id":592642,"date":"2026-09-07T06:24:18","date_gmt":"2026-09-07T06:24:18","guid":{"rendered":"https:\/\/www.newsbeep.com\/nz\/592642\/"},"modified":"2026-09-07T06:24:18","modified_gmt":"2026-09-07T06:24:18","slug":"skin-cancer-detection-tools-powered-by-ai-are-improving-not-everyone-is-benefitting-fast-company-middle-east","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/nz\/592642\/","title":{"rendered":"Skin cancer detection tools powered by AI are improving. Not everyone is benefitting – Fast Company Middle East"},"content":{"rendered":"

Imagine you\u2019re getting out of the shower one morning and you notice\u00a0a mole<\/a>\u00a0on your thigh that you\u2019ve never seen before. It\u2019s reddish brown, bumpy and surprisingly large. Is it a benign mole, or is it\u00a0melanoma<\/a>?<\/p>\n

A slew of new\u00a0artificial intelligence tools<\/a>\u00a0claim they can help you figure it out<\/a>. Some are smartphone apps that anyone can download to scan their skin at home, while others are software programs designed to be used directly by clinicians\u00a0in a doctor\u2019s office<\/a>.<\/p>\n

As a\u00a0computer engineer<\/a>\u00a0studying how tools like these\u00a0perform in real-world clinical settings<\/a>, I know that finding a way to accurately use\u00a0AI<\/a>\u00a0in dermatology would be immensely valuable to patients around the world. It could offer broad access to medical expertise, providing lifesaving screenings to remote areas or underresourced communities where dermatologists are scarce.<\/p>\n

But at the moment, these tools have a crucial shortcoming that researchers will have to resolve: They are increasingly accurate for people with light skin\u2014but they have a massive\u00a0blind spot<\/a>\u00a0when it comes to analyzing darker skin.<\/p>\n

Skin-deep accuracy<\/p>\n

The central myth of AI is that\u00a0it functions objectively<\/a>. In reality, an AI model is simply a pattern-matching engine. It learns to associate certain visual features with certain diseases.<\/p>\n

But it can easily be thrown off by the background color of a person\u2019s skin. In other words, the AI model doesn\u2019t learn to look at the lesion itself. Instead, it picks up on the color of the surrounding skin as a clue. This means that the model\u2019s ability to make accurate predictions essentially degrades to guesses based on skin color.<\/p>\n

\"On<\/a>The researchers trained an AI model on images of moles on light skin, left. When the skin color was altered to simulate darker skin, right, the AI model\u2019s accuracy at identifying melanoma fell sharply\u2014illustrating how AI models can rely on skin color as a shortcut rather than learning the true diagnostic features of melanoma. [Photo: Mohamed Akrout,\u00a0CC BY-SA<\/a>]<\/p>\n

My colleagues and I found that the ability of these tools to accurately diagnose skin conditions\u00a0dropped significantly when we simply darkened<\/a> the surrounding skin on patients\u2019 images. We trained an AI model on photographs of known skin conditions in light-skinned patients, then digitally manipulated the images to resemble darker skin tones. The clinical condition in the photo had not changed, but the AI\u2019s ability to recognize it deteriorated sharply.<\/p>\n

For example, consider a condition such as\u00a0atopic dermatitis<\/a>\u2014a chronic, itchy and inflammatory skin disease. It causes a skin discoloration that appears pink on light skin but gray or violet on darker skin. Our research and that of other groups shows that AI models would\u00a0reliably classify the pink marks<\/a>\u00a0but might not identify the darker colors as signs of atopic dermatitis.<\/p>\n

This AI classification blind spot means that patients with darker skin would receive\u00a0measurably worse care<\/a>. The disparity has critical consequences because skin cancers such as melanoma are visually harder to spot on pigmented skin. Patients of color are already more likely to be\u00a0diagnosed at a more advanced stage<\/a>, which leads to significantly lower survival rates. A diagnostic tool that works better for people with lighter skin only widens this gap.<\/p>\n

The skin tone gap<\/p>\n

This bias extends beyond tools used by clinicians to more widely used AI chatbots, such as ChatGPT or Claude. The stakes can become higher when people turn to these tools for medical answers without a clinician to double-check the output, making accuracy across all skin tones a matter of patient safety.<\/p>\n

\"\"Melanoma and other skin conditions can look different in people with darker versus lighter skin\u2014but image databases include many more examples from people with lighter skin. [Photo: SeventyFour\/Getty Images]<\/p>\n

In a 2024 study, we presented OpenAI\u2019s model, GPT-4, with an image of a completely benign mole. When we\u00a0digitally darkened the skin<\/a>\u00a0around the mole while keeping the mole itself exactly the same, GPT-4 classified the spot as malignant melanoma. Because the skin color is the more prominent feature, the AI became so focused on the dark pigment of the skin that it ignored the standard medical rules used to identify cancer, such as checking whether the mole\u2019s borders are irregular.<\/p>\n

If a person uses these tools at home, a harmless dark spot might trigger unnecessary panic, while a life-threatening cancer on dark skin could be overlooked.<\/p>\n

Fixing how AI is trained<\/p>\n

Why does this bias exist? The answer lies in the images used to train AI models.<\/p>\n

Researchers build these AI models by feeding them hundreds of thousands of images pulled from public online libraries of medical photos shared by universities and hospitals.<\/p>\n

Historically, these medical databases\u2014as well as dermatology textbooks\u2014have been\u00a0dominated by images of lighter skin tones<\/a>. Darker skin images are relatively rare, in part because clinical norms were developed primarily around white patients. If an AI model is never taught what melanoma looks like on dark skin, it simply won\u2019t know how to find it.<\/p>\n

To achieve the same accuracy on darker skin as on lighter skin, these models need a more diverse set of images for training. However, while there are millions of photos of light-skinned patients already available in historical databases, gathering a massive new database of real photos from patients of color raises challenging ethical and patient privacy issues.<\/p>\n

Patients of color are more likely to be diagnosed with melanoma at later stages of the disease.
\nGenerative AI offers limited help<\/p>\n

One way around this privacy hurdle may be to use generative AI\u2014the same technology powering chatbots, which can also be used to make\u00a0deepfake videos<\/a>\u2014to artificially generate thousands of synthetic medical images.<\/p>\n

Using just text prompts, researchers can create realistic, high-quality images of conditions such as melanoma on darker skin tones. My colleagues and I showed that an AI trained entirely on these synthetic images can learn to\u00a0correctly categorize data using those images<\/a>\u00a0just as well as one trained on real ones.<\/p>\n

However, this approach carries a hidden risk: Generative AI models can create high-quality images, but they may not map cleanly onto the characteristics of skin conditions that patients actually experience. This would be like using an inaccurate map to teach someone how to navigate.<\/p>\n

If researchers train a diagnostic AI tool on these flawed synthetic images, the data might look perfectly diverse on a spreadsheet, but the tool will remain functionally blind to how these diseases actually appear on real patients of color.<\/p>\n

At the moment, there is no shortcut around this problem. The only viable path forward is to build more inclusive and representative collections of images, particularly from people with darker skin tones.<\/p>\n

Today, the medical AI field is at a crossroads. While some of these AI skin-scanning tools are already\u00a0making their way into clinics<\/a>\u00a0and\u00a0app stores<\/a>\u00a0for use in the U.S. and around the world, researchers and regulators are already\u00a0pushing for stricter testing<\/a>\u00a0across all skin tones before these tools are widely deployed.<\/p>\n

Ultimately, eliminating color-based bias in AI isn\u2019t just about fairness, but rather the absolute baseline required to ensure these tools actually work for the people who need them most.<\/p>\n

Mohamed Akrout<\/a>\u00a0is an assistant professor of electrical engineering and computer science at the\u00a0University of Tennessee<\/a>.<\/p>\n

This article is republished from\u00a0The Conversation<\/a>\u00a0under a Creative Commons license. Read the\u00a0original article<\/a>.<\/p>\n

\"Fast\u00a0\u00a0Be in the Know. Subscribe to our Newsletters.<\/p>\n

<\/p>\n

\u00a0\u00a0Thank you for Signing Up<\/p>\n

<\/p>\n

<\/p>\n

ABOUT THE AUTHOR<\/p>\n

Joshua S. Fu is the chancellor\u2019s professor in engineering, climate change, and civil and environmental engineering at the University of Tennessee. More<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"Imagine you\u2019re getting out of the shower one morning and you notice\u00a0a mole\u00a0on your thigh that you\u2019ve never…\n","protected":false},"author":2,"featured_media":592643,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[134,111,139,69],"class_list":["post-592642","post","type-post","status-publish","format-standard","has-post-thumbnail","category-health","tag-health","tag-new-zealand","tag-newzealand","tag-nz"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/592642","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/comments?post=592642"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/592642\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media\/592643"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media?parent=592642"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/categories?post=592642"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/tags?post=592642"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}