Could Humans Exist Outside the Loop?

There’s a common expression, especially in the context of AI—“human in the loop.” This refers to human feedback given to otherwise autonomous systems that complete a task. Humans are added as an extra measure of precaution, to verify the outcome, or to help the system improve through feedback. It’s such a widespread pattern that there are even on-demand services, like the aptly named [https://humansintheloop.org\](https://humansintheloop.org), that provide human data labeling to anyone interested. There is no question that in any somewhat critical or semi-critical use of AI, there should be a human in the loop to evaluate the outcome (prediction or generated content) and help with model retraining through feedback.

But can we imagine a case where humans could be completely removed from the process? Is it possible to have autonomous processes not initiated, controlled, or verified—and sometimes all of the above—by humans? As technology becomes more deeply ingrained in everyday processes and is designed to accomplish even more complicated tasks, we tend not to see that human in the loop, even if it continues to play such a central role.