At the recent HIMSS Global Health Conference & Exhibition in Orlando, I gave a talk focused on protecting against some of the pitfalls of artificial intelligence in healthcare.
The aim was to encourage healthcare professionals to think deeply about the realities of AI transformation, while providing real-world examples of how to proceed safely and effectively. My goal was to have everyone in the audience break through the hype with me and focus on a mature understanding of how to build this exciting future.
Thankfully, my message was well received. Participants appreciated the potential that emerges when you move beyond the fear of gimmicks and missing out. This represents a higher level of leadership, where thoughtful individuals collaborate across departments to establish clear and actionable goals to improve outcomes.
The appetite for this post-hype approach to AI was so great that I felt it necessary to write a brief summary of the talk and share it widely with Healthcare IT News readers.
I'll briefly touch on the AI time bomb that has already exploded, provide 10 tips for avoiding it, and share two examples of organizations I work with that are implementing AI correctly.
Something you can not do
Hastily launched AI efforts, both inside and outside the healthcare sector, are already showing signs of failure.
For example, Air Canada's customer-facing chatbot falsely promised passengers discounted flights. The company then tried to claim it wasn't their fault, claiming that AI was a separate legal entity “responsible for its own actions.” Unsurprisingly, Canadian courts have not accepted the “It wasn't us, it was the AI” defense, and airlines are now obligated to honor the falsely promised discounts.
Last year, the National Eating Disorders Association planned to replace its experienced helpline staff with Tessa, a chatbot designed to assist individuals seeking advice about eating disorders. But just days before Tessa's scheduled release date, it turns out the bot started offering questionable advice, including recommendations to limit calorie intake, weigh yourself frequently, and set strict weight loss goals. Did. Although Tessa never went live, the incident highlighted the potentially devastating consequences of deploying AI solutions too quickly.
A recent paper published in JAMA Open Network highlights multiple instances of biased algorithms perpetuating “racial and ethnic disparities in health and healthcare.” The authors detailed several instances of biased and harmful algorithms that have been developed and deployed that negatively impact “access to or eligibility for interventions and services, and the allocation of resources.”
And it's especially worrying because many of these biased algorithms are still in operation.
Simply put, the AI time bomb has already exploded and will continue to do so unless proactive steps are taken to mitigate these issues.
what will you do
To help leaders address the risks associated with AI, I've created 10 tips for tackling AI transformation in a safe and sustainable way. These tips are designed to ensure healthcare executives achieve the highest possible return on their investment.
-
Prioritize transparency and explainability. Choose an AI system that provides transparent algorithms and explainable results.
-
Implement robust data governance. Ensuring high quality, diverse, and accurately labeled data is critical.
-
Engage early with ethics and regulatory agencies. Understanding and aligning ethical guidelines and regulatory requirements early can prevent costly revisions and ensure patient safety.
-
Foster interdisciplinary collaboration. The multidisciplinary approach ensures that the AI tools developed are practical, ethical, and patient-centered.
-
Ensure scalability and interoperability. AI tools should be designed to integrate seamlessly with existing health IT systems and be scalable across different departments and organizations.
-
Invest in continuing education and training. Investing in ongoing education and training will enable your staff to effectively use AI, interpret its output, and make informed decisions.
-
Develop a patient-centered approach. Adopt AI practices that enhance patient engagement, personalize care delivery, and avoid inadvertently exacerbating health disparities.
-
Continuously monitor performance and impact. Develop employee and patient feedback mechanisms to enable continuous improvement of AI tools to better meet stakeholder needs.
-
Establish a clear accountability framework. Clearly define the scope of responsibility for AI-assisted decisions.
-
Promote an ethical AI culture. Encourage debate on the ethics of AI, promote the responsible use of AI, and ensure that decisions are made with the welfare of all stakeholders in mind.
Use these tips to help you with your AI journey. Use these to develop principles, policies, procedures, and protocols to get your AI working correctly the first time and to deftly navigate your instances when things don't go as planned. By proactively adopting these tips early in your AI transformation, you can save time, money, and ultimately lives.
what others are doing
AI transformation requires multiple foundational components working together. As I mentioned in my HIMSS talk, like a Thanksgiving rite of passage, we are moving from the AI kid's table where conversations center around ChatGPT to the adult table where leaders are proactively promoting initiatives. It's time to move on. The foundation for a mature AI transformation.
Two of these key elements that I have focused on when working with large healthcare organizations are taking a holistic approach to deployment and investing in a robust data-driven culture.
A healthcare system developed a blueprint for safely implementing large-scale language models. This blueprint covers a variety of impact areas to consider, including the financial and privacy implications of LLMs, and includes important questions to ask in each of these areas.
The objective was to pose specific and interconnected questions to all executives about the risks and benefits associated with LLM implementation. This approach helps highlight trade-offs, such as speed and safety and quality and cost, and provides this diverse group of leaders with a common language to identify opportunities and discuss risks.
Another health system developed 10 key performance indicators to ensure leaders, teams, and processes all contributed to a data-driven, AI-enabled healthcare culture. We also created a survey based on these KPIs to establish a baseline understanding of where our data culture is strong and where it could be improved.
By focusing on understanding clinicians' data needs and delivering high-quality, relevant data when needed, the organization is rapidly improving “good numbers” such as employee engagement and patient satisfaction. And we achieved a remarkable increase.
This is a great example of how AI transformation begins well before the emergence of emerging technologies and hype. By focusing on fundamentals like data, leaders can win quickly while setting their organizations up for lasting success.
What's next?
The future of healthcare requires a “leadership first, technology last” mindset. Managers must prioritize not only the needs of their employees, but also the challenges and opportunities inherent in the process.
This approach involves using science to understand organizations in a systematic and predictable way and relying on high-quality data to generate accurate and reliable insights to guide change. .
Adopting a leadership-first, technology-second mindset also means that decision makers combine science and data with hard-won experience to craft solutions tailored to specific situations. .
This is why the American Medical Association defines AI as “augmented intelligence,” emphasizing its role in augmenting human intelligence rather than replacing it. Their definition emphasizes the importance of keeping our cognitive and emotional abilities at the forefront of decision-making before turning to technology.
Business leaders who embrace these timeless human traits will foster a mature, AI-powered future.
Dr. Brian R. Spisak is an independent consultant focused on digital transformation in healthcare. He is also a fellow at the National Preparedness Leadership Initiative at the Harvard T.H. Chan School of Public Health, a faculty member at the American College of Healthcare Executives, and the author of the following books: calculative leadership.
