Andy Beck Of PathAI On The Future Of Artificial Intelligence: An Interview With David Leichner
Andy Beck of PathAI: AI will continue to get better over time. Systems are continually improving and will be considerably more advanced looking ahead just five years from now. This expectation of progress gives us in the AI industry optimism when we’re running into difficulty in some areas.
As part of our series about the future of Artificial Intelligence, I had the pleasure of interviewing Andy Beck.
Andy Beck, the Co-founder and CEO of PathAI, which is on a mission to improve patient outcomes with AI-powered pathology. Andy earned his MD from Brown Medical School and completed residency and fellowship training in Anatomic Pathology and Molecular Genetic Pathology from Stanford University. He completed a PhD in Biomedical Informatics from Stanford University, where he developed one of the first machine-learning based systems for cancer pathology. He’s been certified by the American Board of Pathology in Anatomic Pathology and Molecular Genetic Pathology and has published over 110 papers in the fields of cancer biology, cancer pathology, and biomedical informatics.
Thank you so much for joining us in this interview series! Can you share with us the ‘backstory” of how you decided to pursue this career path in AI?
Beginning my career as a pathologist and medical researcher at Harvard Medical School, I was both frustrated and emboldened by the limits of pathology, and the opportunity to reach a new phase in disease understanding and therapy identification. Traditionally a manual study, pathology relies on the trained eye to analyze slide images, and yet the wealth of data contained in tissue samples is far more than could ever be measured with human observation alone. This inability to fully unlock the vast information in tissue samples is a barrier to accurately matching candidates to the right therapies.
Advances in AI around 2012–2013 inspired me to explore how to harness AI to boost the depth, breadth and quality of slide image viewing. While at Harvard Medical School and MIT, my colleagues and I trained a machine learning system that showed strong performance for the identification of metastatic breast cancer. Given how well this technology was already performing, we could see that with further development, this approach could revolutionize pathology and transform the diagnosis of disease. It was at this point that we left academia to found PathAI with the mission to improve patient outcomes with AI-powered pathology.
What lessons can others learn from your story?
The most exciting and innovative work often happens at the intersection of different fields. By combining expertise and knowledge from multiple domains, in our case pathology and AI, you can create something unique. Although two industries may seem very different at first glance, I urge people to explore the synergies between different areas of expertise, as it could lead to groundbreaking discoveries and innovations. In a similar vein, it is really powerful to build a team of experts who span differing areas but united by a common mission. By combining even modest expertise in diverse fields, teams can be formed that bring fresh and unique perspectives and can make truly impactful contributions.
Can you tell our readers about the most interesting projects you are working on now?
Cancer and non-alcoholic fatty liver disease (NAFLD) are two examples of areas where PathAI is doing interesting work and seeing some great results. NAFLD is a growing public health concern and today affects about a quarter of Americans with prevalence growing both in the US and abroad. The most advanced form of NAFLD is called NASH and is a major cause of liver failure. Despite the prevalence of NASH, there are currently no approved therapies available. And in the past several years, there have been many failed NASH drug development programs that were unable to hit their drug efficacy endpoints in clinical trials. Interestingly, the reason that some of these programs have failed may not necessarily be because the drug is ineffective. Research by PathAI and NASH pharma partners has shown that, when we go back and look at trial data from certain failed trials and re-do analysis with AI, the drug in fact showed positive signal when pathology is analyzed with AI. This suggests that it was human variability in NASH disease scoring that unfortunately negatively affected the results. Now through our technology, we are working with NASH drug developers to deliver AI-based NASH scoring tools that will more accurately and robustly measure drug effect in their clinical trials so we can finally get effective therapies to patients.
On the cancer side, we’re working with pharma to address the fact that, in general, the majority of patients with advanced cancer still do not respond to treatment even if they present with the targeted biomarker. Further, many patients who initially show response to treatment go on to develop resistance. We as a research community still don’t know why this happens, but what we do know is that many of these answers will be found in pathology. But not through traditional human pathology but by applying AI to analyze the cellular make-up of tumors to see exactly how cells and tissues in the tumor and the tumor micro-environment are impacting drug resistance. At PathAI, we have created many algorithms that are quantifying the cells and features in tumors that can potentially lead to novel, AI-only biomarkers to predict response.
None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards who helped get you to where you are? Can you share a story about that?
When I was focused solely on pathology, I had a wonderful mentor — Daphne Koller, a computer scientist with a concentration in computer vision and artificial intelligence. I’m so grateful that she took a chance on me, professionally, as I wasn’t the type of person she usually worked with. The opportunity to learn and work alongside Daphne allowed me to make some of the earliest progress in machine learning for pathology and helped me get my start in this industry at a pivotal moment.
What are the 5 things that most excite you about the AI industry? Why?
1. The impacts of AI are far-reaching. AI’s impact can transform essentially everything that humans are currently trying to do today. Since the different applications of AI are incredibly broad, it’s going to be up to us (humans) to effectively utilize the technology to ask the most important questions that can only be answered through this technology.
2. AI can be widely distributed. Connected devices are everywhere these days. For example, there are smartphones in the pockets of billions of people around the world and cloud computing covers the globe. Before the age of smart technology, expertise was always confined to a person and could only be shared in very limited terms. Now, once a problem is solved with AI, it will be very simple to bring that technology to billions of people due to the ease of access.
3. AI will continue to get better over time. Systems are continually improving and will be considerably more advanced looking ahead just five years from now. This expectation of progress gives us in the AI industry optimism when we’re running into difficulty in some areas.
4. AI is attracting the best people. AI is going to be so important for everyday life and this observation is creating somewhat of a snowball effect that is attracting the brightest minds to the field. The current level of innovation and the quality of people who are thinking about how AI can revolutionize our lives is really exciting.
5. The rate of progress in AI is incredible. Within the last 10 years, AI has improved considerably in terms of speech and image recognition. Today, AI programs are generating human-sounding text. Nearly every year, we’re seeing big advances in the space and how it’s being commoditized and utilized. To me, AI is the technology that has seen the most rapid speed of improvement in the past 10 years.
What are the 5 things that concern you about the AI industry? Why?
1. Poor communication. I worry about AI being overhyped and expectations being missed in certain applications due to miscommunication. The term AI itself is super broad and often not clearly used to communicate what the technology is capable of. This can lead to a misunderstanding of what AI can and can’t do today versus what it can and can’t do in the future. This could be alleviated by providing more clarity on what a specific product does if it is powered by AI.
2. Risk of training on biased datasets. All training datasets are biased in some ways, and there is a risk that AI systems trained on biased datasets could actually make those biases even worse by increasing their scale in the real world. AI ethics are really important and taking into account the potential negative impact of different types of biases in the training of AI systems. It’s critical for the field to understand and be transparent about both what the potential positive and potential negative impact of deploying an AI system could be.
3. Over-trust on AI systems. The explosion of AI being accessed by everyday people could lead to people losing valuable skills by becoming overly reliant on the work being performed by AI. At that point, it could become more difficult for humans to identify and correct errors in these systems. So, we need to be vigilant to continue to monitor these systems carefully, to be very skeptical of their performance, and to keep focusing on continuous improvement.
4. AI should focus on the most impactful applications. Experts working with AI should put their energy into developing programs that benefit humanity and the greater good of the world. It concerns me when efforts are invested in AI applications that do not have clear positive benefits (like making social media apps leveraging powerful AI to make their apps more addictive).
As you know, there is an ongoing debate between prominent scientists, (personified as a debate between Elon Musk and Mark Zuckerberg,) about whether advanced AI has the future potential to pose a danger to humanity. What is your position about this?
AI absolutely has tremendous potential for advancing humanity but also poses significant risks. This is typical in that any system might potentially not work out as planned. That’s not to say that AI and robots are going to take over the world, but rather that there can be unintended consequences of AI. A real-world example of this is self-driving cars causing fatal accidents.
There are significant risks of AI misuse and, as the technology becomes more powerful and society relies on it more and more, those risks will only increase. One of our biggest responsibilities in the AI industry is maximizing the benefit while minimizing the risks. This includes developing best practices for AI programs and continued research.
The benefits that AI can bring to making pathology faster and more accurate far outweigh the negatives, but people should still be aware there’s a risk the program could give the wrong diagnosis — just like there’s a risk they can get the wrong diagnosis today without the use of AI.
What can be done to prevent such concerns from materializing? And what can be done to assure the public that there is nothing to be concerned about?
First, we need to encourage a much deeper understanding of the strengths and limitations of AI-based systems. I think a key to reducing this risk is a world where everyone should be educated in AI — even starting with our children. Given AI will be permeated in everything, just as we teach children fundamentals of writing/grammar and mathematics in school, we need to be teaching AI and statistical concepts to everyone, not to just those who decide to major in computer science in college.
Beyond that, for specific applications there are certain fundamental essential elements for all AI systems, including training on large and varied data sets, validation of algorithms on datasets representative of the real world, and working closely with regulators to ensure the safety and quality of systems, and lastly, clearly communicating the benefits and potential risks to the public are vital ways to mitigate harm.
How have you used your success to bring goodness to the world? Can you share a story?
One story I can share is actually someone else’s story that illustrates what AI will do. We have had the honor of working with Wayne Eskridge who leads the Fatty Liver Foundation. Wayne’s story starts even before PathAI. Wayne’s initial liver biopsy reported no NASH when in fact, he was positive for the disease. His diagnosis was not corrected until years later, and his experience led him to found the FLF to increase awareness among NASH patients about the challenges with diagnosing and scoring NASH. Now that we are bringing AI tools to solve this issue, there is a world where the better accuracy in diagnosis from AI will enable physicians to better serve patients, but it will also better empower patients to take charge of their disease and make the best decisions.
As you know, there are not that many women in your industry. Can you advise what is needed to engage more women into the AI industry?
There’s a lot to be learned from really listening to women in the industry to understand what either attracts or drives them away from pursuing more opportunities in the AI field. PathAI has hired women into senior leadership roles that did not come from AI backgrounds but had valuable expertise in other areas, which they’ve subsequently applied to immersing themselves in PathAI and our technology. I encourage AI companies to be proactive in terms of who they reach out to and the way they conduct interviews, and to invest in cultivating a very supportive and inclusive corporate culture that people want to be a part of. Leaders should be proactive about hiring, supporting and promoting great women leaders who will then become role models for a broader group.
What is your favorite “Life Lesson Quote”? Can you share a story of how that had relevance to your own life?
My favorite life lesson quote is essentially: “Whatever has the nature to arise will also pass away.”
This quote has relevance to me every day. I find it motivating and it enables me to remember the preciousness of each day and that the only thing we can count on is change.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂
At PathAI, our company mission is to improve patient outcomes with AI-powered pathology. I would love to start a movement to really focus on solving inequalities in healthcare around the world (which reach far beyond pathology). For AI-powered pathology to truly fulfill its full potential, it will need to be coupled with a much larger and broader movement to bring the best in preventive medicine, access to diagnostic procedures, and access to therapies to areas of the world with far too few physicians and healthcare infrastructure. In the context of this larger movement, AI-powered pathology could fully support patients around the world by enabling not only deployment of the best diagnostic algorithms, but also access to healthcare professionals and treatment to take action on those diagnoses to prevent and treat disease, and to reduce morbidity and mortality.
How can our readers further follow your work online?
LinkedIn: https://www.linkedin.com/in/andrew-beck-a77328aa/
Thank you so much for joining us. This was very inspirational, and we wish you continued success in your important work.
About The Interviewer: David Leichner is a veteran of the Israeli high-tech industry with significant experience in the areas of cyber and security, enterprise software and communications. At Cybellum, a leading provider of Product Security Lifecycle Management, David is responsible for creating and executing the marketing strategy and managing the global marketing team that forms the foundation for Cybellum’s product and market penetration. Prior to Cybellum, David was CMO at SQream and VP Sales and Marketing at endpoint protection vendor, Cynet. David is a member of the Board of Trustees of the Jerusalem Technology College. He holds a BA in Information Systems Management and an MBA in International Business from the City University of New York.
Source: Medium
