Artificial intelligence is infiltrating healthcare at multiple levels. However, discussions around building vs. buying, what kind of impact it may actually have and how expensive expertise is are dominating the space.
Here are 37 impressions of AI in healthcare.
The executives featured in this article are all speaking at Becker's 9th Annual Health IT + Digital Health + RCM Meeting: The Future of Business and Clinical Technologies which will take place Oct. 1-4, 2024, at the Hyatt Regency in Chicago.
If you would like to join this event as a speaker, contact Randi Haseman at rhaseman@beckershealthcare.com.
As part of an ongoing series, Becker's is talking to healthcare leaders who will speak at our conference. The following are answers from our speakers at the event.
Question: What is your impression of AI in healthcare?
Tom Andriola. Chief digital officer and vice chancellor for information, technology and data of UC Irvine: I think the question really should be, 'What is your impression of AI in healthcare so far?' I see the impact of AI as a long-game, general purpose technology. In that regard, think of the Tour de France with 21 stages, and we are in Stage 2. Early AI leaders are trying to establish themselves as viable competitors while the incumbents are trying to stay close enough to the leaders to not lose relevance.
In organizational discussions around AI, it seems prudent to view it as a general purpose technology that will dramatically affect all knowledge-related work over the coming decade. In the short term, we’ll see benefits in the form of knowledge augmentation, incremental productivity improvement and hopefully some burnout relief. However, healthcare will start taking greater interest with the emergence of AI agents and multi-modal AI that can deliver better engagement and outcomes. In the longer term, organizations will have to grapple with a greater step back to rewire its thinking and redesign workflows and processes pretty dramatically or else they're likely to fall behind. Startups will not just be in terms of delivering innovative products and services to the healthcare market, but we’re likely to see innovative models that are 'AI-at-the-center,' which create new delivery models for patients in more personalized and streamlined ways. By the end of the decade, I think we’ll be able to see which organizations have embraced AI and successfully incorporated it into their operations and patient facing strategies. By 2035, I hope we’re telling stories that start with 'remember when,' where we can’t imagine how we worked without these tools.
Tina Esposito. Senior vice president and chief data officer of Advocate Health (Charlotte, N.C.): AI in healthcare will be revolutionary, but the work ahead is not simply how quickly we can implement it. The focus must be how best to responsibly harness the power of AI for short-term benefit yet considering the long-term implications are not fully understood.
Reid Stephan. Vice president and CIO of St. Luke's Health System (Kansas City, Mo.): Today, when people talk about AI in healthcare, they are mainly talking about generative AI in healthcare. Generative AI in healthcare represents a potentially significant leap forward in personalizing patient care and streamlining medical processes. However, its integration must prioritize putting the human – both patient and provider – at the center of its application, ensuring that technology enhances rather than replaces the human touch in caregiving.
Michael Pfeffer, MD. CIO of Stanford Health Care and associate dean and clinical professor of medicine of Stanford School of Medicine (Palo Alto, Calif.): AI has the potential to revolutionize patient care by augmenting the work of our highly trained clinical care teams. AI can streamline administrative tasks, improve operational efficiency, ease cognitive burden for overtaxed clinicians and, ultimately, improve lives by enabling more accurate and timely interventions. However, careful consideration must be given to data privacy, ethical concerns and ensuring that AI technologies complement, rather than replace, human expertise and compassion in healthcare delivery. The RAISE (Responsible AI for Safe and Equitable) Health initiative at Stanford Medicine is focused on ensuring that AI in healthcare must be fair, useful and reliable.
Michelle Stansbury. Vice president for innovation and IT applications of Houston Methodist: AI is driving healthcare innovation, especially when it’s being used to cater to the patient experience to improve accuracy, safety and efficiency. At Houston Methodist, we’re utilizing AI to support both patients and clinicians in a number of ways; one of which is our virtual operations center. This center uses AI to enhance processes and helps guide clinical decision making across our healthcare system. The program was implemented during COVID to address staffing challenges and now the virtual operations center has grown even more, and works collaboratively with bedside teams to capture trends in real time and includes telesitting, telenursing and remote monitoring using biosensor technology.
We’re also using AI technology for specific-use cases in areas including imaging AI as well as in operating rooms to accurately track OR procedures which will help improve OR case durations, accuracy with start times, turnover time spent and surgical instruments used. These are just a handful of examples of how AI technology is helping to streamline workflows, and Houston Methodist continues to explore and evolve how we’re using AI within our healthcare system to deliver quality, compassionate care for our patients.
Tahir Ali. Chief technology and chief information security officer of Montage Health (Monterey, Calif.): AI has the potential to revolutionize healthcare by improving patient outcomes, reducing costs and increasing efficiency. However, it is crucial that we carefully monitor and verify the work of AI systems to ensure their accuracy and effectiveness. Human oversight and intervention will remain essential as we continue to integrate AI into the healthcare industry.
Benjamin A. Hohmuth. Chief medical informatics officer of Geisinger (Danville, Pa.): We are probably at peak hype for AI in healthcare, but the potential is enormous, and the plateau of productivity will be high. I think over the next year we’ll see maturation and scaling of certain use cases like ambient documentation that will be transformative for many clinicians. I also suspect we’ll start to see some use cases fall short of expectations. I wouldn’t be surprised if draft responses to patient messages fell into this bucket. I think integrating AI more smoothly into native workflows will be an ongoing area of opportunity, and, of course, we’ll see a proliferation of new applications. I’m very optimistic about the future of AI in healthcare!
Sha Edathumparampil. Corporate vice president and chief data officer of Baptist Health South Florida (Coral Gables, Fla.): According to studies, administrative tasks account for between 15-30% of healthcare spending in the US. The integration of intelligent automation, which combines AI and automation technologies, can significantly improve efficiency. This advancement promises to streamline operations, minimize redundancies and allow healthcare professionals to devote more time to patient care. It's easy to imagine a near-future scenario where almost every administrative task within hospital systems is supported or even fully managed by AI assistants. Ideally, a process enhanced by AI would be redesigned to proactively address some of today's most common healthcare challenges. For example, an AI-powered virtual scribe can transcribe consultations and review a patient’s medical history to ensure that visit summaries and documentation accurately reflect the correct codes and conditions. This could lead to a decrease in claim denials and a more streamlined revenue cycle. The synergy between redesigned processes and AI assistants has the potential to generate a self-reinforcing cycle that continuously improves healthcare delivery and administrative efficiency.
Areas of clinical and provider assistance, such as disease diagnosis and personalized treatments, are rapidly adopting AI, especially in fields like imaging and oncology. Although widespread adoption may take time, these advancements are already helping systems enhance patient care and achieve better health outcomes.
Last but not least, patient experience stands to benefit significantly from the adoption of AI. Use cases, including personalized patient experiences that reduce friction throughout the care lifecycle like smart appointment scheduling systems, are set to become increasingly common. These systems may predict the length of appointments based on the reason for visit and historical data, ensuring that sufficient time is allocated, helping reduce wait times and improving patient satisfaction.
Jeffrey M. Hoffman, MD. Chief medical information officer of Nationwide Children's Hospital (Columbus, Ohio): AI is showing tremendous promise for many use cases in healthcare. The most immediate tangible benefits are coming from efficiency improvements in operational processes, especially mundane and repetitive administrative tasks. There will also likely be positive impacts from more intelligent and useful chatbots for scheduling and revenue cycle workflows. In terms of direct patient care use cases, like clinical chatbots, note generation and summarization tools, creation and translation of patient educational materials, etc., I’m only cautiously optimistic at this point. There are still significant concerns about accuracy, hallucinations and patient privacy to fully embrace the technology yet, but many are working hard to overcome these barriers, and I anticipate a lot of progress in the coming years.
Louis Jeansonne, MD. Chief medical information officer of Ochsner Health (New Orleans): While there is no substitute for a physician's clinical judgment, many of the tasks that take up a great deal of physicians' time lend themselves very well to AI assistance. Reviewing and summarizing charts, generating notes and addressing in basket messages – the time it takes to do these are already being reduced by large language models. This will allow physicians to focus more on their patients and to have a more manageable workload.
Lanie L. Dixon. Vice president of patient experience of Essentia Health (Duluth, Minn.): I believe that incorporating AI into healthcare should prioritize augmenting human capabilities rather than replacing them. Human-AI collaboration can lead to better patient outcomes, improved workflow efficiency and enhanced healthcare delivery. Maintaining the human touch ensures that patient care remains personalized, compassionate and ethical amidst technological advancements.
AI indeed has the potential to enhance efficiency and effectiveness in healthcare processes, such as diagnosis, treatment planning and administrative tasks. However, maintaining the human touch and ensuring that human judgment remains central is crucial to providing compassionate care, ethical decision making, trust, communication, contextual and cultural understanding, accountability and oversight.
Keith Duemling. Director for cybersecurity technology protection of Cleveland Clinic: AI will clearly be an essential part of healthcare in the future like it will be an essential part of many aspects of our lives. But I think we’re still on the upswing of the hype cycle – I don’t think we’ve peaked in our inflated expectations.
At my own institution, we’re still working on getting good governance into place. We are using AI tools in some areas. For example, we are using Microsoft/Nuance DAX Copilot to produce notes for clinic visits. The tool is pretty amazing in general, but I don’t think it’s quite living up to the hype consistently. Some providers love it, but others find that they still need to spend significant time editing the notes. For example, it still has issues in getting the patient's gender correct. Some of these providers are just going back to dictating their notes which also uses speech-to-text technology, not transcriptionists. No doubt the Copilot product will improve substantially over time, but I do see it as an example of where the reality doesn’t quite yet live up to the expectations.
Zafar Chaudry, MD. Senior vice president, chief digital officer and CIO of Seattle Children's: The impression of AI in healthcare is a mixed bag. On the positive side, it's seen as a powerful tool with the potential to improve diagnoses, personalize treatment and streamline administrative tasks. This could lead to better patient outcomes and a more efficient healthcare system. However, there are concerns. People are worried that AI might replace human interaction in medicine, leading to a loss of the crucial patient-doctor connection. Additionally, the lack of transparency in some AI algorithms raises questions about accountability and potential biases. Addressing ethical concerns and ensuring human oversight are crucial for successful implementation.
Laura Prietula. Deputy CIO at Department of Veterans Affairs: AI has existed since the 50s, so it is not new. What is different is when, where and how pervasive we use it. We use AI in multiple solutions, products and services in various sectors. These uses are transforming lives, processes and ecosystems. Healthcare is benefiting from the application of AI.
AI algorithms can now analyze huge amounts of structured and unstructured data quickly. In healthcare, AI has enhanced clinical decision-making by providing clinicians with real-time insights, evidence-based recommendations and predictive analytics. AI-driven tools, used responsively, have streamlined clinical workflows, automated routine tasks and optimized resource allocation, which lead to improved clinical efficiencies, reduced errors and better patient outcomes.
AI-powered technologies, such as telehealth, have extended healthcare services to underserved populations, rural areas and regions with limited resources, improving access to quality care. Such technologies have also bridged gaps in healthcare delivery and empowered patients to manage their health actively.
I am diabetic, and while my continuous glucose monitoring system is not AI-powered (yet), the 'app' I use with it has algorithms to help me monitor glucose levels, predict patterns and help me change to more healthy eating behaviors.
As we become hungrier for AI uses in healthcare, we must focus on patient well-being and autonomy (help them make better health and healthcare choices independently). While continuing to identify ethical considerations, understand privacy concerns, eliminate algorithmic bias, evaluate security risks and ensure equitable access to AI-powered healthcare solutions.
There are 'AI summers,' when there is a rapid growth of AI-related activity, and 'AI winters,' where the pace slows down. We may be in a long-lasting and perhaps never-ending 'AI spring.' AI has become a 'household' name, and its value is still being discovered. We have much to explore about the use of AI in healthcare; as a geek and a nerd, I thoroughly enjoy new and creative ways of leveraging AI.
Ebrahim Barkoudah, MD. System chief and regional chief medical officer of Baystate Health (Springfield, Mass.): In my opinion, the intersection of AI and healthcare represents a crucial juncture for the advancement of medical care, offering innovative means to enhance the efficiency and effectiveness of patient services. The potential benefits of integrating AI into healthcare are considerable: encompassing task automation, enhanced analysis of extensive patient data for improved care and significant cost reductions. Hybrid chatbots, for instance, promise to reduce the workload of healthcare professionals, cut down on unnecessary procedures, and improve patient education and adherence to treatments. The scope of AI applications within healthcare is vast, ranging from rudimentary administrative tasks to sophisticated aspects of medical diagnostics, treatment planning and the creation of new therapeutic approaches. However, this transition is not without its risks. A significant concern is the potential for AI to erode the personal connection between patients and healthcare providers—a fundamental component of effective healthcare. Additionally, there is the risk of bias in AI algorithms, which can inadvertently perpetuate existing health disparities if not carefully addressed. Rushing the integration of AI into healthcare settings without adequately considering these aspects could lead to trust issues or ethical dilemmas. Therefore, while the move towards AI-enhanced healthcare holds immense promise, it necessitates a balanced approach that maintains the human element in patient care and ensures ethical standards and fairness.
Crystal Broj. Chief digital transformation officer of Medical University of South Carolina (Charleston, S.C.): In the world of healthcare, AI's making some pretty interesting moves. We've got three main players on the field. First, there are the bots and automation tools. They are basically the heavy lifters taking care of the routine stuff, like scheduling appointments or pushing papers digitally. Healthcare's getting pretty cozy with these guys since they make life easier. Then, we have the natural language AI – think voice assistants and chatbots. They're super promising for better communication, but honestly, the healthcare world's still trying to figure out the best way to use them and help their staff adopt them. The trickiest of the bunch is the prescriptive AI; the kind that dives into heaps of data to fish out insights, maybe even help diagnose. While it sounds like a game-changer, getting there is tough because it's all about having the right data and knowing what to do with it. It’s an exciting time but healthcare and AI are still a work in progress, with some parts moving smoother than others.
Alexandra Natale Jackson. Director for digital transformation of Virtua Health (Marlton, N.J.): AI in healthcare shows considerable potential for automating repetitive tasks, granting access to extensive clinical data and complementing human capabilities in areas such as tumor detection. This presents an opportunity to diminish administrative burdens, optimize workflows and dedicate additional resources to focus on direct patient care. However, it's crucial to recognize that although AI holds immense promise, it isn't a one-size-fits-all solution for every healthcare challenge.
Kris Seymour. Director for transformation and project management operations of WellStar Health System (Marietta, Ga.): The current state of AI in healthcare RCM reveals a gap between possibility and practice. We're not fully capitalizing on its potential.
While buzzwords like 'AI' dominate discussions, the implementation of effective AI solutions across healthcare RCM remains limited. This underutilization hinders the ability to address critical challenges. We need a wider adoption of AI-powered tools to streamline tasks and optimize workflows. Tasks that can be significantly improved with wider AI adoption are ones like:
- Automating repetitive tasks: Streamline claim scrubbing and submission processes with AI to minimize errors and expedite reimbursements.
- Proactive claim management: Utilize AI to analyze vast datasets and predict potential claim denials before submission. This allows for preemptive corrections and significantly reduces revenue leakage.
- Intelligent patient engagement: Implement AI-powered chatbots to efficiently answer patient billing inquiries, offer flexible payment options and guide them through the process. This improves patient experience and accelerates payment collections.
By extending the reach of AI across various RCM functions, we can empower staff. Repetitive tasks become automated, freeing up valuable time for healthcare professionals to focus on complex cases, patient communication and strategic RCM. This translates to improved operational efficiency, reduced costs and, ultimately, a more sustainable financial landscape for our patients.
Therefore, it's crucial to move beyond the hype and actively embrace the practical application of AI in RCM on a wider scale. This strategic utilization presents a significant opportunity to optimize processes, minimize revenue leakage and ensure the financial health of the healthcare system.
Carol Yarbrough. Business operations manager for Telehealth Resource Center of UCSF Health System (San Francisco): My impression is that AI, like any diagnostic or other tool used in medicine, will be valuable in assisting providers to arrive at informed medical decision-making. It will not supplant a provider’s judgment but, like an x-ray or lab result, create an amalgam of the results for commonalities for consideration.
Garrett Olin, MBA. CIO of Shasta Community Health Center (Redding, Calif.): I think that AI has started making inroads into healthcare in many areas. It has shown how effective it can be in its current state and shows great promise for the future as AI matures. The accuracy of results will continue to improve and become an integral assistant in all functional areas. Great efficiencies and automation will also improve the overall experience for healthcare professionals, patients and patient care.
Bryan Graven. CIO and executive director information services of Eastern (Manchester) Connecticut Health Network: The impact of AI in healthcare is profound, spanning from improved diagnostics and treatment planning to streamlining administrative tasks and enhancing patient outcomes. AI can help with early disease detection, personalized medicine, predictive analytics and resource optimization. The thought is that AI will lead to a more efficient and effective healthcare delivery network. While all of this sounds great, ethical, privacy and regulatory considerations need to be evaluated. That said, the potential is endless and will have a tremendous potential to revolutionize healthcare by making it more accurate, available and patient-focused. It will also change the resource dynamics between healthcare providers and patients and how we utilize care teams today.
Saima Aftab, MD, MBA. Chief strategy officer and vice president for business development of Nicklaus Children's Health System (Miami): In the future, AI will revolutionize healthcare: optimizing diagnostics, personalizing treatment plans and improving patient care through advanced imaging analysis and predictive modeling. Virtual health assistants powered by AI will enhance accessibility and provide real-time support, easing the burden on healthcare providers and improving efficiency. However, challenges such as data privacy concerns, algorithmic bias and the potential displacement of healthcare workers must be addressed. Ethical oversight is crucial to mitigate risks and ensure that AI integration remains patient-centric and equitable. Balancing the transformative potential of AI with ethical considerations will be essential for realizing its full benefits in healthcare delivery.
Nadim Ilbawi, MD. Physician director for innovation and care models of Endeavor Health (Chicago): I believe AI has the potential to revolutionize healthcare in a manner we have not seen for decades. This will require building trust on both the patient and clinician end. As such, the first forays into this new horizon should focus on elements of care that both reduce the administrative burden on clinicians and improve the patient experience without compromising safe, quality care.
Stephen DelRossi, MSA. CFO and Interim CEO of Northern Inyo Healthcare District (Bishop, Calif.): I am very excited for AI in healthcare. As the market is very early in its development, I must move slowly and deliberately, only choosing products with minimally invasive implementations with the most cost-effective profile. Cost containment is always reviewed, and decisions are made based on the profile. I also focus on the maturation of the product and where I am looking to have the biggest impact. I believe many products have not reached the first layer of maturation, so I will wait and see if improvements are made or if other vendors have better designed products. To that end, I have brought on two products designed to help with the revenue cycle documentation and coding. As they were brought in over the last couple of weeks, I have not decided if we will retain the product at the end of the lease, but I am very excited about the potential it holds.
C. William Hanson III, MD. Vice president and chief medical informatics officer of Penn Medicine (Philadelphia): AI will be a huge disrupter in healthcare across a variety of domains including patient care and communications, clinician documentation, administrative and billing functions, scheduling and patient progression. Our job as healthcare professionals is to make sure we use AI as a force for good, avoiding the dangers of bias or undue reliance on the tool.
Chris DeFlitch, MD, FACEP, FAAEM. Vice president, chief medical information officer of Penn State Health, professor of Penn State College of Medicine and emergency physician of Milton S. Hershey Medical Center (Hershey, Pa.): There has been a history of various tools for clinical decision support over the years, and I think that machine learning and AI are additional assets to advance CDS. Additionally, generative AI has some potential to ease some of the administrative burden. Our early operation use of generative AI tools suggest current focused areas of success with a bright future toward more expanded use. However, just like the early days of the EMR, there is lots of possibility but no defined standards. We don’t want to see overregulation or over-administration, like forcing use of AI for ALL medical decision-making. However, as the field advances, it could be an exciting time to see where AI can improve knowledge and patient-physician experience.
Judith Wolfe, MD. Enterprise associate chief experience officer and academic emergency physician of Cleveland Clinic: AI in healthcare has revolutionized patient feedback analysis. Rather than drowning in the sea of survey verbatims, AI swiftly deciphers sentiments and extracts key themes, unlocking valuable insights for healthcare providers. This means understanding patient needs on a deeper level, identifying emerging trends and pinpointing areas for improvement with precision.
ChatGPT's natural language processing capabilities empower healthcare professionals to grasp the nuances of patient feedback effortlessly. It's like having an expert translator for the patient's voice, ensuring that no valuable insight goes unnoticed. This streamlined process not only enhances patient care strategies but also fosters a culture of continuous improvement within healthcare organizations.
At the Cleveland Clinic, we're dedicated to enhancing the patient experience across all service lines. We gauge ease to get care through custom survey questions and employ cutting-edge ChatGPT technology to distill insights from patient feedback. This information forms a crucial part of our comprehensive performance evaluation. Local teams leverage the Cleveland Clinic improvement model to swiftly identify areas for enhancement, crafting action plans with precise tasks, milestones and a built-in PDCA process. It's not just about listening to our patients – it's about taking concrete steps to elevate their care experience every step of the way.
Ashley Toney. Director for revenue cycle and chief compliance and privacy officer of St. Luke's Hospital, Atrium Health (Charlotte, N.C.): In most instances, AI can be and has been a positive step for healthcare. We all are facing issues every day or see opportunities that AI can assist with. The patient should be top of mind when considering AI. Now more than ever, safeguarding information is key as well.
However, to a large amount of healthcare, AI is inaccessible based on available funding. Some AI is a nicety, and some is a necessity. Unfortunately, the necessity is becoming more prevalent and remains a tricky situation for smaller facilities to remain relevant and properly serve their community.
Kari Nelson. Director for consumer centricity of St. Charles Health System (Bend, Ore.): My impression of AI in healthcare is that we need to continue to keep our patient's experience at the top of our minds. If AI can provide a safer and better experience for the patient and offload low value work from caregivers, then we should adopt it and embrace it. Staffing shortages continue to be a challenge, and patients want a more consumer-centric, on-demand experience in healthcare. If AI can help us solve and overcome these barriers, then we should be adopting its capabilities.
Michael Archuleta. CIO and HIPAA and information security officer of Mt. San Rafael Hospital (Trinidad, Colo.): I firmly believe that AI represents a beacon of hope in the realm of healthcare delivery. Its potential spans from predictive analytics to personalized medicine, offering avenues to enhance clinical decision-making and streamline administrative processes, thereby fostering improved patient outcomes. The transformative power of AI shines particularly bright for rural hospitals, which grapple with unique challenges such as limited resources and access to specialized expertise. Through features like remote monitoring, telemedicine and decision-support systems, AI stands poised to bridge these gaps, empowering rural providers to elevate the quality of care and overcome geographic barriers.
Moreover, AI-driven analytics can revolutionize resource allocation and bolster operational efficiency in rural hospital settings, ensuring sustainability even in the face of economic constraints. Nonetheless, the successful integration of AI necessitates meticulous attention to ethical, regulatory and privacy considerations, ensuring that it complements, rather than supplants, human expertise and compassion. This thoughtful approach will be critical in maximizing the benefits of AI while upholding the integrity of healthcare delivery.
Brad Hoyt, MD. Chief medical information officer of Ardent Health Services (Nashville, Tenn.): These tools are mutually beneficial for patients and healthcare professionals. As we continue to navigate a labor shortage, new technologies must improve employee and provider workflows and experience. Generative AI and machine learning can help reduce burdens on those team members and bolster clinicians’ toolboxes for delivering care and improving outcomes.
Reed Smith. Chief consumer officer of Ardent Health Services (Nashville, Tenn.): AI in healthcare has real promise. Initial focuses should be on supplementing the workforce, easing burdens and enhancing efficiency. There are great opportunities to harness AI for task-based endeavors like summarizing and translation that help pave the way for streamlined processes. However, the next frontier lies in leveraging AI for clinical aids such as charting, providing second opinions, predictive analytics and ensuring compliance, marking the next significant stride in revolutionizing healthcare.
Laura Groschen. Executive vice president and CIO of Acadia Healthcare (Franklin, Tenn.): AI in healthcare has the potential to transform the way we diagnose, treat and prevent diseases. It can improve the quality and efficiency of healthcare delivery, reduce costs and errors, and enhance patient outcomes and satisfaction. AI can also enable new discoveries and innovations in medicine, such as personalized treatments and precision health.
However, AI alone will not solve all the challenges and limitations of healthcare. AI is a tool that needs to be used responsibly, ethically and transparently by human professionals who understand its strengths and weaknesses. AI should not replace human judgment, compassion and accountability, but rather augment and complement them.
Raymond Lowe. Senior vice president and CIO of AltaMed (Commerce, Calif.): Whether we call it augmented intelligence or artificial intelligence, AI is here to stay. I’m intrigued and look forward to several AI opportunities, including ambient clinical voice and NLP, large language models for clinical decision support and patient engagement. As IT professionals and leaders anticipating the future of AI in healthcare, we need to ensure that we have defined governance, business cases and proper oversight. We cannot fully trust AI as there are known cases of AI hallucinations when information has been reviewed. Also, we have to stay vigilant to ensure there is no bias in these technologies that would create or foster health equity issues for our patients and communities.
Thomas M. Maddox, MD, MSc. Vice president for digital products and innovation of BJC HealthCare (St. Louis) and Washington University School of Medicine: AI in healthcare is exciting in concept and untested in implementation. As both an innovator and a clinician, I’m focused on trying to realize AI’s possibilities while bringing the appropriate scrutiny, evaluation and rigor to its deployment. Although I’m not concerned that AI will eventually replace doctors, I’m certain the doctors using AI will, and should, replace doctors not using AI.
Khang Nguyen, MD. Assistant executive medical director for care transformation of Southern California Permanente Medical Group (Pasadena, Calif.): AI offers a tremendous opportunity to improve patient care while reducing a physician’s administrative tasks, so they can reconnect with their patients. For example, AI and machine learning have shown promise in early warning systems that analyze hospital patients’ data to identify when patients are at risk of serious decline and may need intervention. Computer vision technology using AI can analyze medical images for tumors, cancers and surgical guidance. We can also leverage data, telehealth, AI and machine learning to reduce asynchronous workflow, such as charting in the EHR, managing the in-basket, etc. The integration of AI into RPM also holds great potential for enhancing patient care – our teams are actively using AI to analyze data from wearable devices to anticipate and mitigate health issues before they escalate.
While there is excitement about generative AI being able to learn after being given a huge amount of data, there is also concern. Errors can propagate, so we must be cautious as AI can have unintended consequences. That’s why we like to call AI 'augmented intelligence' because we believe the human element simply cannot be taken out of the loop.
Edward Lee, MD. Director for clinical informatics of California Northstate University College of Medicine: There has never been more excitement in AI than now, and I don't see that letting up anytime soon. The question is how much of the excitement will translate into tangible benefits for our patients and health systems. Some AI tools are ready now while others will still take additional time to mature. Either way, the pace of change will be more rapid than we've ever seen before.
Karyn Baum, MD, MSEd, MHA, FHM. Adjunct professor of medicine and adjunct professor for the School of Public Health of University of Minnesota (Minneapolis): AI has stunning possibilities in healthcare, from supporting care operations to improving precision medicine to every clinician having an AI assistant, thus allowing providers to move away from the computer terminal and back to our patients. Having AI reach its potential will require the industry to innovate and rapidly experiment, in a safe manner, with this technology. Just as we did during COVID, we need to become comfortable moving forward with the best available information and rapidly altering as superior techniques and information become available.