New research from MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) suggests that AI itself might be able to accurately adjudicate whether a task is best suited to a human or to technology.
The researchers have developed a machine learning system to assess a task and identify whether it’s one that would be best performed by a human expert or technology.
This appeared to be borne out by the MIT team, whose human-AI hybrid model was able to perform 8% better on both tasks than either human or technology could on their own.
It’s noticeable that the rapid adoption of health technologies during the Covid-19 pandemic has largely seen (relatively) new technologies bolted onto existing processes.
Rather than a face-to-face consultation, for instance, they are done instead over telehealth platforms.
“One of the most telling messages from healthcare provider leaders was that the implementation of new AI technology was not about the solution per se, but about how it was implemented and the success of the change process around its introduction.
AI will take on some of the tasks it is well suited to, freeing up the highly skilled doctors to do the work they are best suited to.
Developments such as that highlighted by the MIT team are a fascinating indication of the progress being made, but it’s clear that there is a long way to go before such technologies are a mainstream part of healthcare as we know it.