1. Providers are using unauthorized AI for more efficient workflows
Of those who confirmed they have used unapproved AI tools, 45% said their primary reason was for a faster workflow. An additional 24% said their primary reason was for better functionality than the currently approved tools, supporting the need to include clinician input in enterprise tool selection.
2. Providers are more likely to experiment with unapproved AI tools
One key area where providers and administrators differed was in tool experimentation. Of those who had used unapproved tools, 26% of providers did so out of curiosity or experimentation, compared to 10% of administrators. If enterprise standards aren’t established, clinicians will try out new tools that can suit their workflows.
3. AI tools are used by providers across all lengths of practice
Many organizations are approaching unsanctioned AI tool usage as a challenge primarily with residents and younger clinicians. However, among providers who had less than 5 years of experience, only 14% had used unsanctioned AI tools, and 16% indicated that they had never used AI tools. Of those who had more than 5 years of experience and had used unapproved AI tools, 42% used them for faster workflows, 29% out of curiosity and experimentation, and 23% because they provided better functionality than approved solutions.
An enterprise-wide AI policy is key
For leaders, it’s crucial to clearly establish and communicate AI policies to the entire enterprise. It’s also important for providers to be included in AI tool selection to ensure evidence-based solutions are considered—without input, they may return to using their preferred tools in the shadows.
Explore more survey findings in our free whitepaper, “Shadow AI: A hidden risk for healthcare.”
