Health execs face blind spots with gen AI

 Healthcare executives could be setting themselves up for problems down the line by failing to consider all of the issues involved in generative AI implementation, according to a new survey by Deloitte. 

Leaders may be too narrowly focused on acquiring and protecting enough high quality data to support generative AI without fully considering the governance, usability, and equity issues that surround the real-world deployment of AI tools. 

While the survey was small, with only 60 respondents, it does give some insight into executive priorities as more and more healthcare organizations start to adopt AI tools across clinical and administrative functions. The poll found clear blind spots in three major areas, including the data itself, AI’s impact on the workforce, and especially its ultimate relationships with consumers. 

Is governance already falling by the wayside?

Specialized generative AI models for healthcare-specific applications require huge volumes of data of the highest quality, which can be hard to acquire and curate. It’s no wonder, then, that 82% of respondents list data availability, quality, and reliability as a focus area for their AI efforts. Compliance, privacy, and security are also top of mind for organizational leaders, with just over 70% prioritizing these issues. 

However, they’re not quite as tuned into what happens after the acquisition of their data assets. Only 60% said governance was a major consideration, and a mere 45% flagged mitigating biases as something at the top of the agenda. 

In an environment where the quality, completeness, and inclusiveness of medical and administrative data is still hit-or-miss at best, this could be a big problem for generative AI users.  Organizations will need to increase their investments in standardization, risk assessment, iterative testing, and ongoing monitoring to avoid skewed, unreliable, or even “poisoned” models in the near future. 

Preparing the workforce for an AI-driven future

The AI arms race is anticipated to bring new efficiencies to healthcare organizations, but only if human workers understand and accept the role of AI in their new job descriptions. However, executives might not be doing everything they should to prepare their teams for the influx of generative AI tools. 

Only 63% flagged “reskilling and upskilling” as an area of focus. Just 57% are preparing to manage shifting job roles and workforce demands, and a similar number are actively addressing staff concerns about generative AI. The lack of attention to these areas could result in resistance to using new tools, which diminishes ROI and could lead to loss of talent in a market where organizations can’t afford more vacant positions. 

Leaders will need to clearly communicate the benefits of generative AI and explain how these tools can augment productivity and enhance job satisfaction instead of being a threat. 

Making sure the consumer matters

Perhaps surprisingly, participants in the survey were least likely to prioritize healthcare consumers when considering their AI activities. Only 50% ranked building trust to share data and ensuring equitable access to generative AI solutions as important activities. And just 45% think it’s necessary to focus on patient education about AI and its associated risks.  

This indicates a major disconnect from what patients have been saying since AI burst onto the scene. Consumers have repeatedly expressed concerns about AI producing untrustworthy recommendations or errors that might affect their outcomes, and have been loudly demanding greater transparency and choice around how AI is integrated into their care. 

Taking the time to lay out the pros and cons for consumers before implementing generative AI at scale will be essential for creating an environment where AI is accepted and trusted by patients and their caregivers. Without engaging consumers in the development and deployment of generative AI, healthcare organizations will be unable to realize the full value of these tools. 

The survey concludes by stating that data is important, but it’s not the only thing that matters when jumping into the generative AI ecosystem. Executive leaders will need to broaden the scope of their roadmaps if they are to create an accessible, equitable, efficient future that harnesses the power of generative AI.  

Jennifer Bresnick is a journalist and freelance content creator with a decade of experience in the health IT industry.  Her work has focused on leveraging innovative technology tools to create value, improve health equity, and achieve the promises of the learning health system.  She can be reached at jennifer@inklesscreative.com.