Generative AI 2023
What Are Organizations’ Current Adoption & Future Plans?

In general, the number of organizations adopting generative AI solutions is relatively small, and most who have adopted say it is still too early to report outcomes. Though adoption is particularly low among smaller organizations, many larger organizations and other entities like payers have already embraced various generative AI tools from EHR vendors (e.g., Epic) as well as from Google, Nuance, and OpenAI. The higher adoption among larger organizations is likely attributed to their higher number of resources, which enable them to adopt generative AI solutions more swiftly. In addition, larger organizations often have data more readily available and employ data scientists and specialists who can effectively develop, implement, and drive outcomes with these solutions.

Although only 25% of interviewed respondents have implemented generative AI solutions, 58% say their organization is likely to implement or purchase a solution within the next year. As previously stated, larger organizations are more inclined to invest in a solution than smaller organizations. Those who plan to invest are primarily motivated by the desire to enhance efficiency. Due to the numerous challenges organizations face, including staffing and financial constraints, it is critical for them to experiment with various methods that can drive efficiency across their enterprise.

“The release of ChatGPT sparked a conversation around how we are going to utilize AI in our organization. It has led me to pay a lot more attention to where we are using AI already as a system and to think strategically about how we can use AI to help us do our jobs.” —CMIO

Respondents report several challenges associated with generative AI; the top three challenges are accuracy/reliability, cost/ROI, and security/privacy. Accuracy/reliability is mentioned more than twice as much as the other challenges. Accuracy is critical when using AI in healthcare, and respondents worry that inaccuracies, bias, and AI hallucinations would negatively impact patients and decisions. It is also important for solutions to be proven reliable by users. Cost/ROI is the second most-mentioned challenge. Implementing and maintaining AI infrastructure—including hardware, software, and skilled personnel—can be expensive. Moreover, the initial investment may not yield immediate returns, making it essential for organizations to carefully assess the long-term benefits and potential cost savings of generative AI. This assessment allows organizations to secure funding and support for ongoing AI initiatives; however, proving that generative AI is driving outcomes can be difficult. Security/privacy is another frequently mentioned challenge. Patient data is highly sensitive and subject to strict regulations (e.g., HIPAA), and it is paramount for organizations to protect patient privacy and ensure data security throughout the AI workflow. Organizations must implement robust encryption, access controls, and anonymization techniques to safeguard patient information and maintain compliance with regulatory requirements.