Is the loop your naughty corner?

We weren't invited to the loop, we were sent there. In sync with the acceleration of artificial intelligence (AI), humans began to be referred to as parts within the whole of augmented workflows. Such a position isn't always a serious problem, but it is perplexing.

"Beyond the edge of the world there's a space where emptiness and substance neatly overlap, where past and future form a continuous, endless loop. And, hovering about, there are signs no one has ever read, chords no one has ever heard," Haruki Murakami once wrote¹. Such spiralling, between the splinters of time, seems more logical than humans being confined to loops as supervisors of large language models (LLMs).

The core idea that humans ought to oversee AI is obviously very smart. Being told precisely where we fit in the workflow is what weakens autonomy. Conceptually, it could be said that "the loop" is the naughty corner of an AI-powered process. Not because we misbehave to get there, but because it limits our liberty to lead in a meaningful way.

Think before you let yourself be looped

When you assemble IKEA furniture, there are pictures to help. You know which bits are what and where they are connected. When you're a human in the loop (HITL), it's not always clear which steps are done by what and where they interlock. As Stanford University writes²:

"Human-in-the-Loop refers to AI systems that include human feedback or intervention as part of their operation. In these systems, humans may provide guidance, correct errors, or make final decisions to improve the accuracy and reliability of AI. This approach combines the strengths of both humans and machines, ensuring that complex or high-stakes tasks benefit from human judgment and oversight."

In theory this is fine, but it's also an affixation. In linguistics, affixation³ is the "morphological process in which prefixes, suffixes, or infixes are added to a word root to convey grammatical relationships or alterations in meaning." When humans are put in the loop, they are treated as suffixes to root AI systems. The intent is different, but the "add-on" mentality is mirrored. Human oversight does strengthen AI-powered processes, but one's sense of ownership is symbolically lessened when they're reassigned from in the lead to in the loop.

When a child is sent to time out, they're told to think about what they've done. When we're labelled as HITL, we must think about what the AI is doing and what it's already done. Two totally different scenarios with unexpected similarities. In both instances, the person has not chosen where they've ended up, and that can compromise volition.

Don't follow formulas written by others

In future, children may want to grow up to be humans in the loop. Dream jobs are often shaped by early childhood experiences⁴, and today HITL behaviour is everywhere. Subtle as it seems, seeing ourselves and one another—to the extent that we already do today—as supervisors of generative models is a significant shift in self-perception. This doesn't mean the HITL should be removed in favour of unsupervised automation, it means we may want to consider breaking free of the loop language to reclaim the lead.

Today, people adopt AI willingly and involuntarily; sometimes both. When given generative AI tools specifically—as opposed to predictive systems—individuals can feel obliged to assign large quantities of tasks to computer programs. Consciously doing the work thus becomes supervising (and editing) the outputs of unconscious systems. The human morphs into the HITL via tech-induced transfiguration.

Sending a child to a naughty corner is a mechanism of discipline. Putting a human in the loop is a method of AI governance. Diverse as the scenarios are, they share a significant common thread. Both child and HITL know that they must await further instructions, from either the discipliner or the software system. Like children who don't do what they're told, humans who resist the loop can be seen as rebels, but never rebels without a cause.

Pivot your focus away from programs towards people

Refusing the label of human in the loop doesn't mean disregarding the responsible use of technology. Contrarily, it's a gesture that creates space for deeper consideration on the applications of AI at work and in leisure. Acting as a HITL means moving to an engineered rhythm designed by developers, not people. In some situations, this isn't a worry, but for functions that require empathy and feeling, allowing "AI" to set the pace is problematic.

According to American statistician Xiao-Li Meng⁵, "humans are always in the loop for advancing data science or engineering at all fronts, literally and figuratively." Even so, current research on the HITL principle focuses on "the interaction between humans and machines." This angle deprioritises human empowerment by focusing on programs not people. In the words of Meng: "translating and transforming the advances in data science and engineering into actual value for human societies is a matter of interactions among humans." These conversations can't be held properly in "the loop" that no one can define.

"Our responsibility begins with the power to imagine," Murakami reminds us. "It's just as Yeats said: In dreams begin responsibility. Turn this on its head and you could say that where there's no power to imagine, no responsibility can arise." Like the child liberated from the naughty corner, the human who resists "the loop" doesn't need to be led by engineered instruction. Autonomy is then amplified, allowing individuals to feel empowered during their interactions with tech, not subordinated by software systems. Humans must be freed from prefabricated labels, not looped in systems, to preserve integrity and imagination.

Given the concept's prominence, the title of human in the loop may prevail for some time. But you're the only one who can decide if the HITL label suits your style. After all, invitations to the loop were never sent out, nor were we given the chance to RSVP. How you perceive yourself in relation to technology is nobody's business but your own.