Inventor of Cancer Screening Tool Says AI Could Lead Disease Detection

Imagine a future where a simple infection becomes deadly because antibiotics no longer work. This scenario isn't from a sci-fi novel; it's the grim reality of antimicrobial resistance (AMR) we face today. The financial toll? It's already pushing healthcare expenses up by $66 billion annually, with projections suggesting this could escalate to $159 billion per year by 2050 (if resistance rates follow historical trends and we don't intervene with innovative solutions). Bacterial infections such as Group A Streptococcus (Strep A or GAS), affecting hundreds of millions globally, could soon become a major public health challenge. The inability to diagnose and treat them with timely interventions would have a catastrophic ripple effect, undermining global health security and economic stability. According to a 2023 report by the Center for Global Development, investing in healthcare innovation could yield a remarkable global ROI of 28:1, translating into substantial cost savings, improved patient outcomes, and a more efficient healthcare system. AI plays a key role in this transformation, serving as an assistive tool that delivers fast and reliable testing at the point of care, enabling early diagnosis and timely treatment. To understand how AI is shaping disease detection, when it will go mainstream in healthcare, and whether it could replace human decision-making, I interviewed Peter Whitehead, CEO and President of Light AI. His AI-driven platform detects diseases in seconds using a smartphone camera—starting with Strep A. With over 24 years of experience in health tech innovation, Peter Whitehead is well-known in medical circles for inventing VELscope, a revolutionary and market-dominating oral cancer and disease imaging tool. It was patented in 2000 and has been used in over 50 million oral health examinations by more than 20,000 dental practitioners in 23 countries. In this conversation, we discuss how AI transforms real-world healthcare, the challenges ahead, and what the future holds for AI-based disease detection. Let's dive in. AI in healthcare has seen massive growth, but adoption remains slow. Only about 1 in 10 healthcare professionals currently use AI technologies, while nearly 50% plan to adopt them in the future. What do you see as the biggest barriers preventing AI from becoming a clinical standard? Beyond regulatory and compliance challenges—undoubtedly crucial—there's another equally significant barrier: perception. AI is fundamentally reshaping how healthcare is delivered and accessed, and with that transformation comes uncertainty and fear. Many worry that AI integration will upend entire healthcare systems overnight, making the transition overwhelming. While change is inevitable—requiring training, data transfers, and new security measures—it doesn't have to be disruptive or overly complex. To make AI a clinical standard, we must first build comfort and confidence among healthcare professionals. The key is an introduction through simple yet effective AI tools—solutions that enhance accessibility, improve patient outcomes, and alleviate burdens on medical staff. As an inventor, I've seen firsthand that technology can be both powerful and easy to adopt. With Light AI, I'm proving it again. This is where we must begin. But we cannot ignore the urgency. Every day that AI adoption lags, lives are at stake. Today, nearly 4.5 billion people—more than half the global population—lack access to essential healthcare. AI has the potential to bridge that gap, but only if we integrate it wisely and swiftly. How is AI modernizing disease detection? What key challenges does it solve for healthcare providers, patients, and the broader industry? AI is revolutionizing disease detection and transforming multiple areas of healthcare. By automating time-consuming, high-volume repetitive tasks, AI allows healthcare providers to focus more on patient care rather than administrative burdens. It enhances diagnostic accuracy and speed, reducing dependence on lab tests and swabs that often come with long wait times. At its core, AI is solving some of healthcare's biggest inefficiencies. By streamlining bottlenecked processes, AI improves speed and system-wide efficiency, ultimately allowing healthcare systems to function more fluidly. This will become even more critical as the global population continues to grow. Today, 4.5 billion people lack access to essential healthcare services, and by 2030, a 10 million healthcare worker shortage is projected. Without AI, we cannot bridge these gaps in time. The need for change is urgent—AI isn't just an innovation; it's a necessity. While the benefits are undeniable, regulation is the key to responsible AI integration. We must ensure AI tools are safe, effective, and trustworthy before they become widely adopted. There's an ongoing debate about AI replacing human decision-making in medicine. What makes Light AI an assistive platform rather than a replacement for clinicians? How does it complement medical professionals' expertise while ensuring accuracy and trust in the diagnostic process? We're not looking to replace healthcare professionals. We need them but think about it this way. When a patient openstheir mouth to show their general practitioner the back of their throat, GPs can't say with certainty if something is wrong. If they can see some swelling and tenderness, they cannot be certain about exactly what is causing it. This very scenario, which has played out time and time again, is one of the reasons why today we are dealing with such high levels of antimicrobial resistance. The overprescription of antibiotics "just in case" is not an effective method of care, yet this is the approach that's often taken. In the US, roughly 6 in 10 patients are prescribed antibiotics when only 2 or 3 really need them. Here is where AI can assist doctors in their decision-making process. Light AI's platform truly shines here because inunder a minute, it can detect throat conditions like Strep A, helping physicians diagnose more accurately without having to rely on expensive and time-consuming methods like swabs and lab tests. Our platform has received a 97% accuracy rating in pre-FDA validation studies, and the database we've built contains 280,000+ images of the back of the throat, meaning our AI has a wide and diverse range of data to help it intelligently arrive at a conclusion. AI will not replace doctors, but rather, I believe, people who do not use AI will be replaced. AI-driven diagnostics rely heavily on high-quality data. What are the biggest challenges in collecting, curating, and standardizing healthcare data? How can AI models ensure both accuracy and fairness while avoiding bias? Kim Perry, Chief Growth Officer at Emtelligent addresses this best: "Healthcare has made significant strides in interoperability over the past 15 years, and the CMS final rule regarding interoperability promises to further enhance health information exchange. However, this rule is only the first step." According to Kim, to truly empower patients and place them at the center of care, healthcare data must be not only accessible but also usable. Simply enabling data exchange isn't enough if most healthcare information remains unstructured, fragmented, and locked in incompatible formats. Many organizations struggle to extract clinical and financial value from this data. To fully realize the potential of interoperability, the industry must modernize data processing pipelines and harness AI, natural language processing (NLP), and large language models (LLMs) to make healthcare data actionable. By addressing this last barrier in the clinical data pipeline, healthcare can unlock the full value of digital transformation, improving patient care and outcomes. While the quote above applies to the US market, we need to see this play out in each region to allow for effective data exchange. Patient data is sensitive and subject to strict privacy laws (e.g., HIPAA, GDPR) and often lacks diverse representation, leading to bias in AI models. To ensure accuracy and fairness, AI must be trained on broad, representative datasets that include patients from different demographics, geographies, and socioeconomic backgrounds. Privacy and security are major concerns when it comes to AI in healthcare. With increasing regulatory scrutiny (e.g., HIPAA, GDPR), what steps should AI-driven platforms take to ensure data protection and compliance? AI infrastructure is evolving into a new category supervised by the MLOps function. Typically, this function complies with and participates in the best practices of cybersecurity measures, including prevention from hacking, data breaches, cyberattacks, and other security issues. Essential components of an effective data strategy include multi-key encryption, access controls, and regular security audits. Further, as seen recently, even the use of proprietary models requires authentication to ensure that the applications using them are the ones for which they are intended and authorized. How do you see the balance between private R&D funding and collaboration with public health organizations? How much investment has gone into Light AI's R&D efforts over the past years, and what partnerships has Light AI established to develop its AI-powered platform? The balance is crucial for fostering innovation while ensuring public benefit and accessibility. If you look at the AI healthcare space, you can see there has been and still is a huge appetite for it. As adoption becomes more widespread, innovation in the space will only continue. In terms of Light AI, over $20 million has gone into extensive research and development over the last 8 years to build our platform. We worked with over 13 partners, including UNM Hospitals, one of the largest hospital networks in the world, UCLA Health, The University of Rhode Island, American Heart Association, Uganda Heart Institute, and others, to get to the place we are today. We have three granted patents and one allowed patent on our technology, with plans for FDA approval by late 2025. Your work with Light AI focuses on AI-powered disease detection. What specific conditions does the platform currently detect, and how quickly can it provide results? Our platform is designed to detect Strep A. It is a severe bacterial infection that's reached a 20-year high, according to 2023 preliminary data from the CDC. It's also classified as one of the World Health Organization's top five health threats, and each year, 600 million are infected with StrepA, killing almost as many children as malaria. Strep A was previously known as a relatively harmless bacteria until recently. It's been on the rise, and if it enters into a wound, it can lead to amputations and/or death. Outside of Strep A, we will be looking to expand into other infections like Covid, influenza, and EBV. One of the most common reasons things go south is delayed detection, which we're working to address. Our platform can detect Strep A in about 30 seconds from image collection to result. Whether in person or via a telehealth call, a physician simply has to take an image of the back of the throat with a smartphone and they will receive results in under 60 seconds allowing for quick action to take place if needed. The great thing about the platform is that as long as you have a smartphone, a healthcare professional can reach you and investigate what's going on. That opens up access to approximately 4.5 billion people worldwide. Beyond Strep A, what other diseases or conditions do you see AI-powered diagnostics expanding into? How scalable is the underlying technology for detecting a wider range of conditions? Our platform has been set up to allow for the eventual detection of other conditions, including throat conditions and conditions in the eye and skin. Outside of what we're doing and to my knowledge, AI can detect early signs of more than 1,000 diseases before a patient is even aware of any symptoms which is pretty incredible. It can also spot more bone fractures than humans can and help ambulances determine who needs to be transferred to the hospital and who doesn't. Ultimately, good AI tools require good data. As platforms scale, they will need a sufficient amount of data to accompany their efforts. With the global AI healthcare market expanding, how do you see adoption playing out geographically? Which regions or healthcare systems are leading the way, and where do you see the greatest opportunities for AI-driven diagnostics? The adoption of AI in healthcare is moving at different paces across the globe and is influenced by different factors, including regulation, infrastructure, investment, acceptance, and health care maturity. I would say the regions which areleading include the US, Europe, China, and the Middle East. The regions with newly emerging opportunities include India, Southeast Asia, and Africa. Here, AI-driven diagnostics can address critical gaps in accessibility and affordability, making a truly transformation shift in healthcare for these populations. What experiences shaped your vision for transforming diagnostics? On a personal level, I've always believed that innovation should serve humanity, not just as a concept, but as a tool to bridge gaps and solve real problems. That belief is what drives everything I do. Growing up, I dreamed of becoming a doctor. I wanted to make an impact by helping others. But as I took a closer look, I realized that the problems weren't just within the walls of hospitals—they were in the systems themselves. So many people lacked access to even basic healthcare. The system wasn't just strained; for many, it was inaccessible. I knew I wanted to do something to change that. Looking five years ahead, what's your boldest prediction for AI in healthcare? What breakthrough do you think will fundamentally change the way we approach diagnostics and patient care? This is a bold prediction, but in five years, we could see AI taking on certain healthcare tasks and responsibilities with little to no human intervention. Such a breakthrough, for example, would enable surgeons to concentrate solely on surgeries and allow specialists to focus on the core aspects of their work. However, achieving this will require navigating the complexities of the current regulatory landscape, etc. That said, a lot can change in five years. If this interview resonates with you, please support it by sharing it on your social media and helping spread the word about the transformative potential of AI in healthcare. Imagine a future where a simple infection becomes deadly because antibiotics no longer work. This scenario isn't from a sci-fi novel; it's the grim reality of antimicrobial resistance (AMR) we face today. The financial toll? It's already pushing healthcare expenses up by $66 billion annually, with projections suggesting this could escalate to $159 billion per year by 2050 (if resistance rates follow historical trends and we don't intervene with innovative solutions). pushing healthcare expenses Bacterial infections such as Group A Streptococcus (Strep A or GAS), affecting hundreds of millions globally, could soon become a major public health challenge. The inability to diagnose and treat them with timely interventions would have a catastrophic ripple effect, undermining global health security and economic stability. hundreds of millions According to a 2023 report by the Center for Global Development, investing in healthcare innovation could yield a remarkable global ROI of 28:1, translating into substantial cost savings, improved patient outcomes, and a more efficient healthcare system. AI plays a key role in this transformation, serving as an assistive tool that delivers fast and reliable testing at the point of care, enabling early diagnosis and timely treatment. report To understand how AI is shaping disease detection, when it will go mainstream in healthcare, and whether it could replace human decision-making, I interviewed Peter Whitehead , CEO and President of Light AI . His AI-driven platform detects diseases in seconds using a smartphone camera—starting with Strep A. Peter Whitehead Light AI With over 24 years of experience in health tech innovation, Peter Whitehead is well-known in medical circles for inventing VELscope , a revolutionary and market-dominating oral cancer and disease imaging tool. It was patented in 2000 and has been used in over 50 million oral health examinations by more than 20,000 dental practitioners in 23 countries. VELscope In this conversation, we discuss how AI transforms real-world healthcare, the challenges ahead, and what the future holds for AI-based disease detection. Let's dive in. AI in healthcare has seen massive growth, but adoption remains slow. Only about 1 in 10 healthcare professionals currently use AI technologies, while nearly 50% plan to adopt them in the future. What do you see as the biggest barriers preventing AI from becoming a clinical standard? currently use Beyond regulatory and compliance challenges—undoubtedly crucial—there's another equally significant barrier: perception . perception AI is fundamentally reshaping how healthcare is delivered and accessed, and with that transformation comes uncertainty and fear. Many worry that AI integration will upend entire healthcare systems overnight, making the transition overwhelming. While change is inevitable—requiring training, data transfers, and new security measures—it doesn't have to be disruptive or overly complex. To make AI a clinical standard, we must first build comfort and confidence among healthcare professionals. To make AI a clinical standard, we must first build comfort and confidence among healthcare professionals. The key is an introduction through simple yet effective AI tools—solutions that enhance accessibility, improve patient outcomes, and alleviate burdens on medical staff. As an inventor, I've seen firsthand that technology can be both powerful and easy to adopt. With Light AI, I'm proving it again. This is where we must begin. But we cannot ignore the urgency. Every day that AI adoption lags, lives are at stake. Today, nearly 4.5 billion people—more than half the global population—lack access to essential healthcare. AI has the potential to bridge that gap, but only if we integrate it wisely and swiftly. How is AI modernizing disease detection? What key challenges does it solve for healthcare providers, patients, and the broader industry? AI is revolutionizing disease detection and transforming multiple areas of healthcare. By automating time-consuming, high-volume repetitive tasks, AI allows healthcare providers to focus more on patient care rather than administrative burdens. It enhances diagnostic accuracy and speed, reducing dependence on lab tests and swabs that often come with long wait times. At its core, AI is solving some of healthcare's biggest inefficiencies. By streamlining bottlenecked processes, AI improves speed and system-wide efficiency, ultimately allowing healthcare systems to function more fluidly. At its core, AI is solving some of healthcare's biggest inefficiencies. By streamlining bottlenecked processes, AI improves speed and system-wide efficiency, ultimately allowing healthcare systems to function more fluidly. This will become even more critical as the global population continues to grow. Today, 4.5 billion people lack access to essential healthcare services, and by 2030, a 10 million healthcare worker shortage is projected. Without AI, we cannot bridge these gaps in time. The need for change is urgent—AI isn't just an innovation; it's a necessity. healthcare worker shortage While the benefits are undeniable, regulation is the key to responsible AI integration. We must ensure AI tools are safe, effective, and trustworthy before they become widely adopted. There's an ongoing debate about AI replacing human decision-making in medicine. What makes Light AI an assistive platform rather than a replacement for clinicians? How does it complement medical professionals' expertise while ensuring accuracy and trust in the diagnostic process? We're not looking to replace healthcare professionals. We need them but think about it this way. When a patient openstheir mouth to show their general practitioner the back of their throat, GPs can't say with certainty if something is wrong. If they can see some swelling and tenderness, they cannot be certain about exactly what is causing it. This very scenario, which has played out time and time again, is one of the reasons why today we are dealing with such high levels of antimicrobial resistance. The overprescription of antibiotics "just in case" is not an effective method of care, yet this is the approach that's often taken. In the US, roughly 6 in 10 patients are prescribed antibiotics when only 2 or 3 really need them. Here is where AI can assist doctors in their decision-making process. In the US, roughly 6 in 10 patients are prescribed antibiotics when only 2 or 3 really need them. Here is where AI can assist doctors in their decision-making process. Light AI's platform truly shines here because inunder a minute, it can detect throat conditions like Strep A, helping physicians diagnose more accurately without having to rely on expensive and time-consuming methods like swabs and lab tests. Our platform has received a 97% accuracy rating in pre-FDA validation studies, and the database we've built contains 280,000+ images of the back of the throat, meaning our AI has a wide and diverse range of data to help it intelligently arrive at a conclusion. AI will not replace doctors, but rather, I believe, people who do not use AI will be replaced. AI will not replace doctors, but rather, I believe, people who do not use AI will be replaced. AI-driven diagnostics rely heavily on high-quality data. What are the biggest challenges in collecting, curating, and standardizing healthcare data? How can AI models ensure both accuracy and fairness while avoiding bias? Kim Perry, Chief Growth Officer at Emtelligent addresses this best: " Healthcare has made significant strides in interoperability over the past 15 years, and the CMS final rule regarding interoperability promises to further enhance health information exchange. However, this rule is only the first step. " addresses Healthcare has made significant strides in interoperability over the past 15 years, and the CMS final rule regarding interoperability promises to further enhance health information exchange. However, this rule is only the first step. According to Kim, to truly empower patients and place them at the center of care, healthcare data must be not only accessible but also usable. Simply enabling data exchange isn't enough if most healthcare information remains unstructured, fragmented, and locked in incompatible formats. Many organizations struggle to extract clinical and financial value from this data. To fully realize the potential of interoperability, the industry must modernize data processing pipelines and harness AI, natural language processing (NLP), and large language models (LLMs) to make healthcare data actionable. By addressing this last barrier in the clinical data pipeline, healthcare can unlock the full value of digital transformation, improving patient care and outcomes. large language models (LLMs) While the quote above applies to the US market, we need to see this play out in each region to allow for effective data exchange. Patient data is sensitive and subject to strict privacy laws (e.g., HIPAA, GDPR) and often lacks diverse representation, leading to bias in AI models. To ensure accuracy and fairness, AI must be trained on broad, representative datasets that include patients from different demographics, geographies, and socioeconomic backgrounds. To ensure accuracy and fairness, AI must be trained on broad, representative datasets that include patients from different demographics, geographies, and socioeconomic backgrounds. Privacy and security are major concerns when it comes to AI in healthcare. With increasing regulatory scrutiny (e.g., HIPAA, GDPR), what steps should AI-driven platforms take to ensure data protection and compliance? AI infrastructure is evolving into a new category supervised by the MLOps function. Typically, this function complies with and participates in the best practices of cybersecurity measures, including prevention from hacking, data breaches, cyberattacks, and other security issues. Essential components of an effective data strategy include multi-key encryption, access controls, and regular security audits. Further, as seen recently, even the use of proprietary models requires authentication to ensure that the applications using them are the ones for which they are intended and authorized. How do you see the balance between private R&D funding and collaboration with public health organizations? How much investment has gone into Light AI's R&D efforts over the past years, and what partnerships has Light AI established to develop its AI-powered platform? The balance is crucial for fostering innovation while ensuring public benefit and accessibility. If you look at the AI healthcare space, you can see there has been and still is a huge appetite for it. As adoption becomes more widespread, innovation in the space will only continue. In terms of Light AI, over $20 million has gone into extensive research and development over the last 8 years to build our platform. We worked with over 13 partners, including UNM Hospitals, one of the largest hospital networks in the world, UCLA Health, The University of Rhode Island, American Heart Association, Uganda Heart Institute, and others, to get to the place we are today. We have three granted patents and one allowed patent on our technology, with plans for FDA approval by late 2025. Your work with Light AI focuses on AI-powered disease detection. What specific conditions does the platform currently detect, and how quickly can it provide results? Our platform is designed to detect Strep A. It is a severe bacterial infection that's reached a 20-year high, according to 2023 preliminary data from the CDC. It's also classified as one of the World Health Organization's top five health threats , and each year, 600 million are infected with StrepA, killing almost as many children as malaria. Strep A was previously known as a relatively harmless bacteria until recently. It's been on the rise, and if it enters into a wound, it can lead to amputations and/or death. Outside of Strep A, we will be looking to expand into other infections like Covid, influenza, and EBV. preliminary data top five health threats 600 million One of the most common reasons things go south is delayed detection, which we're working to address. Our platform can detect Strep A in about 30 seconds from image collection to result. Whether in person or via a telehealth call, a physician simply has to take an image of the back of the throat with a smartphone and they will receive results in under 60 seconds allowing for quick action to take place if needed. The great thing about the platform is that as long as you have a smartphone, a healthcare professional can reach you and investigate what's going on. That opens up access to approximately 4.5 billion people worldwide. Beyond Strep A, what other diseases or conditions do you see AI-powered diagnostics expanding into? How scalable is the underlying technology for detecting a wider range of conditions? Our platform has been set up to allow for the eventual detection of other conditions, including throat conditions and conditions in the eye and skin. Outside of what we're doing and to my knowledge, AI can detect early signs of more than 1,000 diseases before a patient is even aware of any symptoms which is pretty incredible. AI can detect early signs of more than 1,000 diseases before a patient is even aware of any symptoms which is pretty incredible. It can also spot more bone fractures than humans can and help ambulances determine who needs to be transferred to the hospital and who doesn't. Ultimately, good AI tools require good data. As platforms scale, they will need a sufficient amount of data to accompany their efforts. With the global AI healthcare market expanding, how do you see adoption playing out geographically? Which regions or healthcare systems are leading the way, and where do you see the greatest opportunities for AI-driven diagnostics? The adoption of AI in healthcare is moving at different paces across the globe and is influenced by different factors, including regulation, infrastructure, investment, acceptance, and health care maturity. I would say the regions which areleading include the US, Europe, China, and the Middle East. The regions with newly emerging opportunities include India, Southeast Asia, and Africa. Here, AI-driven diagnostics can address critical gaps in accessibility and affordability, making a truly transformation shift in healthcare for these populations. What experiences shaped your vision for transforming diagnostics? On a personal level, I've always believed that innovation should serve humanity, not just as a concept, but as a tool to bridge gaps and solve real problems. That belief is what drives everything I do. Growing up, I dreamed of becoming a doctor. I wanted to make an impact by helping others. But as I took a closer look, I realized that the problems weren't just within the walls of hospitals—they were in the systems themselves. the problems weren't just within the walls of hospitals—they were in the systems themselves. So many people lacked access to even basic healthcare. The system wasn't just strained; for many, it was inaccessible. I knew I wanted to do something to change that. Looking five years ahead, what's your boldest prediction for AI in healthcare? What breakthrough do you think will fundamentally change the way we approach diagnostics and patient care? This is a bold prediction, but in five years, we could see AI taking on certain healthcare tasks and responsibilities with little to no human intervention. Such a breakthrough, for example, would enable surgeons to concentrate solely on surgeries and allow specialists to focus on the core aspects of their work. However, achieving this will require navigating the complexities of the current regulatory landscape, etc. That said, a lot can change in five years. If this interview resonates with you, please support it by sharing it on your social media and helping spread the word about the transformative potential of AI in healthcare.