Current and Future AI Uses in Dermatology: An Expert's View
Roxana Daneshjou, MD, PhD, is one of the leading experts on artificial intelligence (AI) in American dermatology. Daneshjou, assistant professor of biomedical data science and dermatology at Stanford University, Stanford, California, leads landmark AI studies, is an associate editor of the journal NEJM AI , and gives presentations about the topic, including one at the recent Society for Investigative Dermatology (SID) 2025 annual meeting where she starkly warned colleagues that 'dermatologists who use AI will replace dermatologists who don't.'
Roxana Daneshjou, MD, PhD
So one could assume that Daneshjou embraces AI in her clinical practice. But she doesn't — not quite yet. While AI is helpful with office tasks that involve writing, she said, it's not currently good enough at handling tasks, such as evaluating skin lesions or helping solve diagnostic riddles.
'You should only use it for tasks where you can easily catch errors and rectify them. You shouldn't use it when you're not sure of the answer or the next step because you could be badly misled,' she said in an interview with Medscape Medical News .
But just wait. 'Eventually, once we have valid, well-validated AI tools that can help with diagnosis and triage, they're going to become essentially standard of care,' Daneshjou said.
The following are excerpts from the interview with Daneshjou about the present and future of AI in dermatology.
What do you mean when you say, 'Dermatologists who use AI will replace dermatologists who don't'?
Daneshjou: That's actually a rehashed phrase originally coined by Curt Langlotz, a radiologist who made the same claim about radiologists. The point is that dermatologists aren't going anywhere. AI is not replacing dermatologists. It's that dermatologists who use AI will replace dermatologists who don't.
Will some dermatologists be left behind?
Daneshjou: Medicine always evolves. There was a time when we didn't have advanced imaging technologies like CT scans and MRIs. And think about how many dermatologists now use electronic health records (EHRs) vs writing everything down by hand. There are still some people writing things by hand, but physicians who can use EHRs have largely replaced those who can't.
This isn't a new phenomenon. Whenever new technology comes along, it becomes incorporated into medical practice, and those who learn to adapt and adopt it eventually replace those who don't.
Is there fear and denial in the dermatology community about AI?
Daneshjou: There's fear, but there's also enthusiasm — sometimes enthusiasm to the point of using things that aren't ready for prime time. In my SID talk, I discussed how it's not safe to use large language models [AI] — LLMs — for any clinical task where you don't know the answer or can't validate it quickly. These models can have errors that are difficult to catch because the outputs look so convincing.
Can you give an example of how using LLMs clinically might get a dermatologist in trouble?
Daneshjou: In my presentation, I showed AI being asked to calculate a RegiSCAR score for a patient. It gives an output that looks really convincing but has some of the scores wrong. If you didn't know the RegiSCAR score yourself, you might not catch that mistake. Similarly, if you ask about medication dosing, sometimes AI gets it right. But research papers show it can get dosing wrong. If you're not certain of the answer, you shouldn't use an LLM for that task.
That's different from giving it bullet points and saying, 'Follow these bullet points to draft a prior authorization letter' or 'Write an after-visit summary for my patient' about a disease you're well-versed in, and you can verify [the text] for accuracy.
Are there reliable clinical uses for AI now?
Daneshjou: First, I should note that publicly facing models aren't Health Insurance Portability and Accountability Act (HIPAA)–compliant, so you have to be careful about putting patient information in them. Some institutions like Stanford have HIPAA-compliant versions internally.
I'd be very wary of using these models for diagnosis and treatment because they can say things that are wrong. I've heard dermatologists say they've put patient images into these models to get a differential diagnosis, which I would strongly advise against — both for HIPAA concerns and because the outputs aren't reliable.
What about 'vision language' models (VLMs) in dermatology that are trained on skin images and could potentially be used for tasks such as identifying lesions?
Daneshjou: The VLMs we've tested perform worse than the LLMs. They're even more in the research realm.
Are current AI systems actually good at categorizing skin lesions?
There are many papers claiming they're good, but there's not much prospective trial data validating that performance. We need more trial data proving that a particular model will continue to perform well in a clinical setting.
So AI isn't ready for prime time in diagnosis and treatment?
Daneshjou: That's correct. It's more useful in a supportive role — helping with writing or editing text.
You worked on a 'red teaming' event that assigned attendees — engineers, computer scientists, and health professionals such as dermatologists to assign medical tasks to AI and ask questions. The results were published in Nature in March 2025. What did you discover?
Daneshjou: We found that across all models tested, there was an error rate of around 20%. As our chief data scientist at Stanford likes to joke, 'You can use large language models for any task where a 20% error rate is acceptable.'
Where do you think AI and dermatology are headed next?
Daneshjou: Image-based models will eventually get good enough to earn US Food and Drug Administration clearance. But my concern is this will happen without the creators having to prove the models work across diverse skin tones — an incredibly important part of validation.
Our research has shown that most image-based AI models exclude diverse skin tones in their training and testing. We're also going to see more multimodal models — ones that incorporate diverse information like images, text, and molecular data — to provide outputs or risk assessments. That's where AI is heading generally, not only just focusing on text or images alone but also taking information from multiple modalities the way humans do.
How often do you use AI in your clinical practice?
Daneshjou: Not very much. I run a research lab, so I use it extensively in research. I've used it to help with grant writing and to analyze recommendation letters I've written, asking it to identify weaknesses so I can improve them. Clinically, I've shown my nurse how to use our secure AI to draft prior authorization letters or rebuttals to insurance [rejections]. But otherwise, I don't really use it in clinic.
You've discussed how AI handles clinical vignettes vs real patients. What should dermatologists understand about this?
Daneshjou: Headlines often misrepresent reality. They'll say, 'AI models can diagnose patients.' But in reality, these models were given very nicely packaged vignettes and were able to provide a diagnosis.
Patients don't come as nicely packaged vignettes. In real clinical practice, I have to ask, 'What's going on?' I have to do the skin check, identify lesions, gather history, and ask about duration, symptoms, occupation, and sun exposure. I have to collect all this information and make a judgment.
Sometimes, the history doesn't match what you see, so you have to use clinical reasoning. This kind of clinical reasoning isn't what they're testing in research papers that claim AI can diagnose patients.
Would you recommend using AI at all for generating differential diagnoses?
Daneshjou: I'm not using AI just to use it. I need a specific reason why I think it will help me. For example, if I'm writing a grant and want a summary of one of my own research papers, I might ask it to write a first draft that I can edit because I know my own research well enough to verify what's correct. But I'm not using it to develop differentials for my patients.
What would you advise dermatologists who want to adapt to AI but don't know where to start?
Daneshjou: The American Academy of Dermatology (AAD) has AI boot camp videos. At the annual AAD meetings, the AAD offers educational sessions on AI.
If you look in the Journal of the American Academy of Dermatology , there are Continuing Medical Education reviews that the AAD's Augmented Intelligence Committee has written to educate dermatologists about AI technologies and what to watch for.
A few years ago, this content was sparse. But there's been a concerted effort to create educational materials for dermatologists.
What would you tell dermatologists who are agonizing about AI?
Daneshjou: I see people posting on LinkedIn what I would call outrageous claims based on research papers. They'll say, 'This research paper shows we have autonomous AI agents that can treat patients,' but when you read the actual paper, it doesn't show that at all. Often, the hype doesn't match the reality on the ground.
And what about those who think AI is overblown and not worth worrying about?
Daneshjou: Claims about AI replacing physicians or dermatologists are indeed overblown. But this is definitely something dermatologists will have to adapt to. It's eventually going to become part of practice in some ways.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Forbes
11 hours ago
- Forbes
The Triple Threat Draining Workforce Productivity And Mental Health
Productivity and mental health are taking a hit for three reasons. Numbers never lie: it's why CEOs obsess over revenue targets, operational metrics, quarterly growth, and more. Yet beneath the spreadsheets and data dashboards lies an invisible crisis quietly eroding organizational strength and productivity: the deteriorating mental health and emotional resilience of today's workforce. A comprehensive 2025 State of the Workforce Report by meQuilibrium, surveying 5,477 employees across various industries, identified a troublesome "triple threat" that undermines employee performance and well-being: uncertainty, pessimism, and psychological disconnection. These aren't mere HR concerns; they're operational risks with quantifiable financial impacts. The data paints a clear picture: These challenges are silently sabotaging productivity, further inflating turnover costs, and increasingly impacting employee well-being. The Real Cost Of The Triple Threat Uncertainty alone led to a 68% increase in productivity impairment. Employees facing high uncertainty stress experience twice the risk of generalized anxiety disorders and nearly double the rate of burnout, costing American companies between $4,000 and $21,000 per employee annually, according to a study in the American Journal of Preventive Medicine. Pessimism compounds this impact significantly. Workers with negative job outlooks report a 64% decrease in productivity, alongside a staggering 128% higher risk of depression and 108% higher risk of anxiety compared to optimistic colleagues. The psychological disconnect further exacerbates productivity losses. Even employees displaying no signs of disconnect report an average productivity impairment of 29%, while severely disconnected employees experience impairment rates of up to 66%. This triple threat isn't fleeting; it's systemic and escalating, making employee mental health and emotional resilience one of your organization's significant, yet often overlooked, financial liabilities. Foundational Strategies To Improve Productivity And Mental Health A company thrives when its people thrive. Amid volatility and continuous disruptions, the greatest threat to any organization is a workforce characterized by stress, negativity, disengagement, and poor health. However, this crisis presents opportunities to cultivate resilience and strength. While each company is unique, these foundational actions can universally elevate employee well-being and productivity: Fitness offers more than aesthetics. It's a strategic investment. CEOs like Bob Iger strongly advocate for fitness because it improves their mood, focus, emotional regulation, executive presence, and resilience. To build a health-driven culture, leaders must actively model wellness practices. Sleep is equally critical. Improved sleep and recovery improve individuals' decision-making and emotional regulation and reduce team volatility. Prioritizing a health-first culture also serves as a strong talent attractor and retention tool. Ambiguity breeds stress, especially with technological disruptions like AI, which create concerns about job stability. Clear and consistent communication significantly mitigates uncertainty. Leaders must frequently reiterate their vision in a way that integrates each person seamlessly. Also key is establishing priorities and performance expectations along with transparency in explaining the rationale behind decisions. Vulnerability and candor strengthen trust, which further reduces speculation and disengagement among employees and investors alike. Soft skills matter significantly, even in the data-driven corporate world. As highlighted in the report, empathetic managers reduce uncertainty stress by 37% and disconnect by nearly 50%. CEOs and senior leaders can actively cultivate empathy through the following: Better Mental Well-Being Equals Improved Productivity Addressing the triple threat isn't merely beneficial—it's critical. The modern workforce faces numerous challenges that threaten employees' mental health and productivity. With ongoing volatility and persistent uncertainty, investing in mental health and resilience becomes not just advantageous but essential. As technology democratizes and ceases to offer unique competitive advantages, the companies that will lead tomorrow are those who invest significantly in their people's well-being today.
Yahoo
15 hours ago
- Yahoo
Pandemic preparedness ‘dramatically eroding' under Trump, experts say
Amid controversial dismissals for independent advisers and staff at health agencies, alongside lackluster responses to the bird flu and measles outbreaks, experts fear the US is now in worse shape to respond to a pandemic than before 2020. H5N1, which has received less attention under the Trump administration than from Biden's team, is not the only influenza virus or even the only variant of bird flu with the potential to spark a pandemic. But a subpar response to the ongoing US outbreak signals a larger issue: America is not ready for whatever pathogen will sweep through next. 'We have not even remotely maintained the level of pandemic preparedness – which needed a lot of work, as we saw from the Covid pandemic,' said Angela Rasmussen, an American virologist at the Vaccine and Infectious Disease Organization at the University of Saskatchewan in Canada. 'But now, we essentially have no pandemic preparedness.' Related: Bird flu reinfections at US poultry farms highlight need for vaccines, experts say 'I'm concerned on a number of fronts,' said Jennifer Nuzzo, professor of epidemiology and director of the Pandemic Center at Brown University School of Public Health. Those concerns include a lack of quality information from officials, weakened virus monitoring systems, and public health reductions at the federal, state and local levels. 'The thing that I am most concerned about is the veracity of information coming out of the health agencies,' Nuzzo said. In the ongoing outbreaks of measles, for example, Robert F Kennedy Jr, the secretary of health and human services, has downplayed the severity of the disease, spread misinformation about the highly effective vaccine to prevent measles, and pushed unproven treatments. 'The communications on measles gives me deep worries about what would happen in a pandemic,' Nuzzo said. 'If a pandemic were to occur today, the only thing we would have to protect ourselves on day one would be information.' The H5N1 outbreak has been plagued by incomplete information, an issue that began in the Biden administration but has amplified under Trump. In Arizona, 6 million chickens were killed or culled at a Hickman's Family Farms location because of H5N1 in May. That's about 95% of the company's hen population in the state. Hundreds of workers, including inmate laborers, are now being dismissed as Arizona braces for egg shortages. We're not testing – it's not that there are no new cases Angela Rasmussen Yet even as H5N1 outbreaks continue to spread on farms and wreak havoc on the food supply, no new bird flu cases have been reported in people for months. 'I am concerned that we may not be finding new infections in humans,' Nuzzo said – and a lack of testing may be the culprit. 'We're not testing – it's not that there are no new cases,' Rassmussen said. The last bird flu case in a person was listed by the US Centers for Disease Control and Prevention (CDC) on 23 February. At that point, at least 830 people in the US had been tested after contact with sick animals. This kind of testing – monitoring the health of people who regularly work with H5N1-infected animals – is how the vast majority of cases (64 out of 70) have been found in this outbreak. But then, several CDC officials overseeing the bird flu response were fired on 1 April. Since then, only about 50 people in the US have been tested after exposure to sick animals – and no positive cases have been announced. It's also been difficult to understand the extent of the outbreak and how the virus spreads among animals. 'We still just don't have a good picture of the scope and scale of this outbreak – we never really have. And until we have that, we're not going to be able to contain it,' Rasmussen said. 'It's extremely bad,' she continued. 'We don't have any information about what's happening right now. The next pandemic could be starting, and we just don't know where that's happening, and we don't have any ability to find out.' We're seeing health departments scrambling. That infrastructure is just dramatically eroding Jennifer Nuzzo Huge reductions in the public health workforce and resources has led to less monitoring of outbreaks, known as disease surveillance. 'Cutting back on that surveillance is leaving us more in the dark,' Nuzzo said. The CDC clawed back $11.4bn in Covid funding in March. This funding was used to monitor, test, vaccinate and otherwise respond to public health issues at the state, local, territorial and tribal level. 'We're seeing health departments scrambling,' Nuzzo said. 'That infrastructure is just dramatically eroding.' International monitoring programs to address outbreaks before they expand across borders have also been cancelled. 'We have taken for granted all of those protections, and I fear that we are poised to see the consequences,' Nuzzo said. Trump's crackdown on immigration also poses a major challenge in detecting cases and treating patients during outbreaks. 'A lot of the people who are most at risk are strongly disincentivized to report any cases, given that many of them are undocumented or are not US citizens,' Rasmussen said. 'Nobody wants to go get tested if they're going to end up in an Ice detention facility.' When cases are not detected, that means patients are not able to access care. Although it's rare for people to become sick with H5N1, for instance – the virus is still primarily an avian, not a human, influenza – this variant of bird flu has a 52% mortality rate globally among people with known infections. Allowing a deadly virus to spread and mutate under the radar has troubling implications for its ability to change into a human influenza without anyone knowing. And if such changes were detected, widening gaps in communication could be the next hurdle for preventing a pandemic, Nuzzo said. 'Communication is our most important public health intervention. People, in order to be able to know how to protect themselves, need to have access to facts, and they need to believe in the messengers. And the communication around the measles outbreaks are deeply eroding our standing with the American people.' Even stockpiled vaccines and other protective measures, like personal protective equipment, take time to distribute, Nuzzo added. 'And flu is a fast-moving disease that could cause a lot of damage in the months it would take to mount a vaccination campaign.' The US government's cancellation of its $766m contract with Moderna to research and develop an H5N1 vaccine also signals a concerning strategy from health officials, Nuzzo and Rasmussen said. Other restrictions on vaccine development, like a new plan to test all vaccines against saline placebos, is 'going to make it extremely difficult to approve any new vaccine' and would 'have a devastating impact on our ability to respond to a potential pandemic', Rasmussen said – especially in a rapidly moving pandemic where speed matters. 'You don't have time for that if this virus causes a human-to–human outbreak,' Rasmussen said. All of these policies mean the US is less prepared for a pandemic than it was in 2020, she said. And it also means there will be preventable suffering now, even before the next big one strikes. 'We are actively making people less safe, less healthy and more dead,' Rasmussen said.
Newsweek
16 hours ago
- Newsweek
Fervo Uses Oil Drilling Technology to Dig Deep for Clean Geothermal Energy
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources. Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content. The U.S. oil and gas industry sparked an energy revolution a few decades ago by thinking in a different direction—literally. Horizontal drilling, combined with hydraulic fracturing, unlocked new petroleum reserves and quickly propelled the U.S. to global leadership in oil and gas. Fervo Energy, a company headquartered in oil capital Houston, is aiming for a similar revolution in geothermal energy by borrowing a page from the oil industry's playbook. "Drilling is drilling," Fervo's Vice President of Strategy Sarah Jewett told Newsweek. "That is kind of the 'aha' moment that our founders really found when they started this business." By using drilling technology advanced by the oil and gas industry, Fervo is drilling deeper, drilling sideways and drilling in new places to unlock new reserves of heat energy. A member of Fervo Energy's drilling rig crew climbs to the top of the rig in Utah. A Fervo well recently reached a depth of more than 15,000 feet, where rock is greater than 500... A member of Fervo Energy's drilling rig crew climbs to the top of the rig in Utah. A Fervo well recently reached a depth of more than 15,000 feet, where rock is greater than 500 degrees Fahrenheit. More Courtesy of Fervo Energy "There's no reason why geothermal energy wells should be using older, more outdated rigs," Jewett said. Traditional geothermal energy relies on finding steam-bearing resources and drilling into them vertically to capture the steam heat. It requires not only hot rock but also enough permeability in the rock to allow water and heat to come together. That made geothermal power geographically limited to those places with just the right combination of factors. Fervo is developing an enhanced geothermal system (EGS) that creates some of those conditions by making rock permeable and injecting liquid where hot rock is available. Instead of just drilling vertical wells, the company uses horizontal drilling and hydraulic fracturing techniques to harvest naturally occurring heat. "We are just using the heat from rock under the earth rather than steam directly to create electricity," Jewett explained. The U.S. Department of Energy calls EGS "the next frontier" for renewable energy. Because the technology can tap a much wider range of geothermal resources, the DOE estimates that EGS has the potential to power more than 65 million American homes and businesses. "Really, hot rock is everywhere," Jewett said. "It's just a question of how deep that rock exists and how economic is it to actually get to [that] depth." Rock that reaches 400 to 500 degrees Fahrenheit is, generally speaking, closer to the surface in the western U.S. than in the eastern states. Fervo is focusing efforts in southwestern Utah, site of the company's Cape Station development where the company is building three power plants. Fervo Energy is building three power plants as part of its Cape Station development in southwestern Utah. The first is scheduled to come online in the summer of 2026. Fervo Energy is building three power plants as part of its Cape Station development in southwestern Utah. The first is scheduled to come online in the summer of 2026. Courtesy of Fervo Energy Jewett said Fervo's first power plant is scheduled to come online next summer with the first phase delivering 100 megawatts of clean power. An additional 400 megawatts of capacity is scheduled at the same site by 2028, and power purchasers include Shell Energy and Southern California Edison. Last week the company announced that a new round of funding for the Cape Station development had secured $206 million from investors including Breakthrough Energy Catalyst, founded by Bill Gates. In another announcement last week, Fervo shared results from one of its exploration wells that pushed the boundaries of geothermal technology. The company's deepest well yet reached more than 15,750 feet, nearly 3 miles beneath the surface. Rock temperature at that depth is greater than 500 degrees, Jewett said, and the company is testing what happens to drill bits, casings and electronics at very high heat. "One of our paths to innovation we call 'boring is beautiful,'" she said. As wells reach greater depth and higher heat, companies providing parts and services test the performance of drilling equipment under increasingly harsh conditions. "That has been a really virtuous cycle of innovation between us and our partners." In another parallel with the oil and gas industry, Fervo had the petroleum consulting company DeGolyer & MacNaughton conduct an independent energy reserve estimate for the Cape Station project. Such reserve estimates are common when oil companies develop a new resource, but the practice is new for the geothermal industry, Jewett said. DeGolyer and MacNaughton applied similar principles from oil analyses to evaluate the thermal energy potential at the site and concluded that Fervo's site could yield 5 gigawatts of power. A gigawatt is enough to power a city of 750,000 homes. With those positive developments on the technical and financial fronts, Fervo is clearly on a hot streak. But like the rest of the clean energy sector, the company is also closely watching policy developments in Washington. Tax credits from the 2022 Inflation Reduction Act were a boon for new clean energy technology developers, not only because of the size of the credits but because the 10-year lifespan of the credits provided long-term stability to attract private investors. The "one big beautiful" bill approved by the Republican-controlled House of Representatives late last month would end most of those clean energy credits and debate is underway on a Senate version. "We're fighting pretty hard to preserve tax credits because for these very early projects, it's incredibly useful to have that that support," Jewett said. Fervo's first phase of development would not be affected by the changes, she said, because construction is already underway, but future projects could become more challenging. Regardless of policy changes, however, she said she is confident that EGS power will find a market as a steady source of electricity to meet surging demand. "Geothermal energy projects are capable of providing that 'always on' power," she said. "And I think that's why it's become so attractive."
