Table 1 Areas of focus for AI and patient safety per RWJF conference.

From: Bending the patient safety curve: how much can AI help?

1.

Develop AI/advanced analytics implementation models, implementation approaches, and methods for integration into clinical workflows

2.

Create a patient safety framework to guide measurement of AI impact: How to use AI to improve each dimension of safety from retrospective analysis to real-time monitoring to future use of prediction

3.

Build an AI patient safety financial business case

4.

Reduce cognitive and total work burden with AI which should be interpretable and usable for frontline users.

5.

AI patient and consumer focused issues: study how patients and consumers will view and use these tools and how their use will impact patient-doctor and patient-healthcare team relationships

6.

Create ways to engage all the relevant stakeholders in AI use and design

7.

Develop effective governance/oversight and accountability for AI in clinical care

8.

Develop methods to learn and loop back to adjust AI algorithms to ensure equity—refine or change for different or changing populations

9.

Create AI to enhance adverse event/near miss monitoring and real time safety surveillance

10.

Create Use Cases for the application of AI to specific problems in patient safety