Healthcare Organizations Prioritize AI Governance to Ensure Ethical and Effective Implementation

According to a recent roundtable session hosted by intelligence platform Elion, 74% of healthcare organizations have established dedicated AI governance committees and frameworks to prioritize use cases.?

These committees play a crucial role in overseeing the development and deployment of AI solutions and ensuring they align with the organization's goals and ethical standards.

Structured governance frameworks allow healthcare organizations to:

• Systematically evaluate AI projects
• Manage risks
• Ensure compliance with regulatory requirements

This foundational step is essential in building trust and fostering a culture of accountability within the organization.

Developing Processes and Policies

While progress continues in establishing governance foundations, more work remains to develop comprehensive processes and policies for adopting and managing AI-driven solutions.?

According to the roundtable, only 50% of healthcare organizations have processes in place for evaluating and monitoring AI products—indicating a significant gap in ensuring ongoing oversight and quality control of AI applications.

Effective AI governance entails an initial assessment of AI solutions, as well as continuous monitoring and evaluation to proactively tackle any emerging issues.

Such ongoing work involves:

• Setting up protocols for regular audits
• Performance evaluations
• Updates to AI systems to ensure effectiveness and safety over time

Managing the Use of Large Language Models (LLMs)

The use of large language models (LLMs) such as ChatGPT is becoming increasingly common in healthcare.?

The roundtable found that 80% of organizations permit the use of LLMs, but there are variations in governance policies regarding their usage. Some organizations allow LLMs only for administrative purposes, while others extend their use to clinical applications within defined scopes. However, some organizations still lack specific guidelines for LLM usage.

This disparity highlights the need for tailored governance policies that reflect each organization's specific needs and capabilities. Clear guidelines and best practices for LLM usage can help mitigate risks associated with these powerful tools, such as data privacy concerns and the potential for generating inaccurate information.

Tailoring Governance to Organizational Context

AI governance strategies are not one-size-fits-all; they must be adapted to the unique context of each healthcare organization.?

Organizational size, technology maturity, and available resources significantly influence governance approaches. For example, academic medical centers (AMCs) often have more in-house AI expertise and may adopt more sophisticated governance frameworks than regional health systems with fewer resources.

Understanding these organizational differences is crucial for developing effective governance strategies. Tailoring AI governance to the specific context ensures that policies are practical, enforceable, and aligned with the organization's capabilities and goals.

Addressing the Need for High-Quality Information and Leadership

Despite the progress in AI governance, there is a strong desire for more high-quality information and thought leadership on key governance topics.?

Healthcare organizations are seeking guidance on several critical areas, including:

• AI literacy, training, and education: Enhancing the understanding of AI technologies among healthcare professionals to ensure they can effectively use and oversee AI applications.

• AI solution life cycle management: Establishing best practices for the entire life cycle of AI solutions, from development and deployment to maintenance and decommissioning.

• Benchmarking for AI solution testing and evaluation metrics: Developing standardized benchmarks and metrics to evaluate the performance and safety of AI solutions consistently.

These areas are pivotal in building a comprehensive AI governance framework that addresses current challenges and prepares organizations for future advancements in AI technology.

In other healthcare AI news ...

What the Progression of AI Automation Means for Healthcare

There are key differences between robotic process automation (RPA), natural language processing (NLP), machine learning (ML), and generative AI. While each contributes to advancing healthcare, understanding their differences is crucial.

Explore key trends and applications

Overcome Clinical Logic Complexities in Telemedicine

Accurately routing patients via telemedicine is still a significant headache for many healthcare organizations due to the complex clinical logic needed for success. Does your organization still face this challenge??

Learn how clinically validated logic can help

Have any questions or thoughts? Let us know in the comments. ??

____________________________________________________________________________

Want to receive the latest digital triage and healthcare AI insights straight in your inbox? Subscribe to our monthly newsletter.?