According to Gartner, the technology research company, by 2021 75% of healthcare organizations will have invested in AI to improve either operational performance or clinical outcomes.

According to Laura Craft, vice president analyst at Gartner, one of the primary needs that level of investment will require is AI governance programs that can help healthcare organizations manage the myriad uses and consequences of AI programs.

In a recent column, Craft worries that “because new AI techniques are largely new territory for most HDOs, there is a lack of common rules, processes, and guidelines for eager entrepreneurs to follow as they design their pilots. Most HDOs have not developed an enterprise strategy for how AI will be introduced, invested in and managed. This leads to a lack of trust in AI-powered solutions and creates a new problem that only healthcare provider CIOs are equipped to address.”

So CIOs, she says, need to take the lead in filling that void with policies and protocols that ensure both discipline and accountability around the use of AI in healthcare.

The good news, Craft says, is that an AI governance body doesn’t have to be separate from every other leadership body within an organization. “If a strategic leadership council for a data and analytics program already exists,” she notes as an example, “then this is the most obvious fit, as AI is a natural extension of an analytics program.”

Regardless of where the governing body is situated within an organization, Craft identifies four “pillars” to successful AI governance: Legal, regulatory and compliance review; Clinical and scientific verification and valuation; Ethical evaluation and usage guidelines, and: Organizational deployment and change management.

More fundamentally, organizations should make sure all their stakeholders, from clinicians to technologists to administrators, are working with a common definition of AI, which will facilitate the development of consistency and thoroughness concerning how best to select and implement AI opportunities.

Finally, Craft notes the obstacle the reality of data can constitute when it comes to the smooth and successful implementation and use of AI. 

“When attempting to curate a clean, complete and accurate dataset,” she says, “(healthcare organizations) are challenged by poor data quality; lack of needed data; incomplete data; and issues of data consent, privacy, and security.”

Consequently, CIOs need to remember to include data governance practices when considering AI governance.