Doctors have started using artificial intelligence in novel ways to communicate with patients and help make diagnoses. Now the government is wrestling with how to ensure the tools do no harm.

Federal regulators are proposing a new labeling system for AI healthcare apps designed to make it easier for clinicians to spot the pitfalls and shortcomings of these tools. The Biden administration has proposed that these apps come with a “nutrition label” that discloses how the app was trained, how it performs, how it should be used and how it shouldn’t.