AI Scribe vs. Human Scribe

Who will win?

Before artificial intelligence, there was human intelligence. To free physicians from the burdens of documentation – and improve their throughput and revenue-generating capabilities – the first crop of scribes were people.

People, however, are people – costly and with their own limitations. Ambient scribe technology has, naturally, exploded as a potential replacement. But, are they “just as good”?

These folks would imply: probably not.

Physicians working with a human scribe spent less time in the EHR – particularly within the notes of the EHR – than those working with the ambient scribe. Note quality, as rated by the PDQI-9 framework was similar, though physicians added a greater number of characters to the ambient scribe notes to bring them up to that standard.

Of further concern, at least for our U.S. friends, encounters featuring ambient scribe use garnered fewer RVUs, on average, than the human scribe:

It is worth noting this little trial occurred back between Dec ‘24 and Jan ‘25, which is an eternity in AI development and integration time. Also, as anyone who has rolled out ambient scribes knows, it can take a significant amount of work to customize the template to fit the specific style and needs of an individual clinician. That said, human scribes are much better at filling in additional context not available to ambient technology, so I suspect humans would remain the winners of subsequent rounds, but the gap would likely narrow.