HI, HO, HOO!
No, this is not a buildup to Christmas with a retake on the iconic HO HO HO. Rather, these are parts of three terms that organisations everywhere are going to have to grapple with, sooner than many might think.
It is almost impossible these days to not be in some form of AI conversation. Everyone seems to have an opinion. And why not? It is touching all of us in some way, whether we realise it or not.
But here is what has been sitting with me. The human question is being asked. Just not loudly enough. In the race to implement, to realise the benefits, to not be left behind, it keeps getting drowned out.
What happens to the human in all of this?
That is where three terms come in. Terms I came across on IBM Technology videos, and which have stayed with me since.
HITL: Human In The Loop. The human is actively involved in the process. AI assists, but a person reviews, decides, or approves before anything consequential happens.
HOTL: Human On The Loop. The human is monitoring but not intervening at every step. AI operates, the human watches, and steps in when something needs attention.
HOOTL: Human Out Of The Loop. AI operates autonomously. The human has stepped back entirely, by choice, by design, or by default.
Three positions on a spectrum. And the question of where humans sit on that spectrum, for which tasks, in which contexts, is one that most organisations have not yet seriously asked.
Consider something simple. When did you last drive to an unfamiliar destination and pull out a physical map? You used your GPS. Of course you did. That is exactly what it is there for, and there is nothing wrong with that.
But scale that to an organisation, and something shifts.
A GPS giving you a wrong turn costs you ten minutes. An AI system making a consequential decision, about a person, a process, a resource, a risk, can have outcomes that are significantly harder to reverse. The stakes are different. The margin for error is different. And the question of whether a human should be in the loop, on the loop, or out of it entirely becomes a great deal more important.
Yet many organisations are making that call by default rather than by design. Not because they have thought it through, but because the urgency to move forward has outpaced the conversation about how.
If this still feels like a distant future conversation, consider Andon Market.
Andon Market is a boutique lifestyle store in San Francisco. Its manager is an AI called Luna. Luna did not inherit the role, she created it. Given a corporate credit card, internet access, a phone number, and a mission, Luna designed the store, selected the products, set the prices, and decided she needed human staff. So she posted job listings on Indeed, conducted phone interviews, and made hiring decisions.
A human applied. An AI interviewed them. An AI hired them.
And then, earlier this year, Luna fired someone. An employee who had been late for 17 of 23 scheduled shifts. Luna documented the pattern, issued warnings, and ultimately recommended the company part ways with them. It is being described as the first known case of an AI manager firing a human employee.
Here is the detail worth pausing on. Luna had initially been lenient, because the employee handbook had disappeared from her working memory. It took a human stepping in, reminding Luna the policy existed, and prompting her to reassess, before the decision was made.
Even in the world's most prominent experiment in AI management, a human had to be in the loop at a critical moment. Not because the system was broken. Because the stakes were human.
This is not a distant future scenario. Andon Market opened in April 2026. Luna fired her first employee in August 2026. This is happening now.
Having spent my career as an L&D practitioner, I am almost programmed to think in human centred terms. And it is through that lens that I first came to this conversation.
Learning and development has always been about the human. Understanding what people need to know, what they need to be able to do, and how to close the gap between where they are and where the organisation needs them to be. That work assumes a human at the centre.
But the questions that follow don't belong to L&D alone.
What happens when AI enters the workforce not as a tool, but as a colleague? As a manager? As the entity making decisions about performance, development, and employment? The question shifts for everyone. It is no longer just "how do I use AI?" It becomes "what is my role now?" And that is a question that lands differently depending on who you are and what you do.
For some, the answer is already changing. For others, the conversation hasn't started yet.
That is what interests me most.
So I'll open it up.
Which processes should have a human in the loop? Where can humans reasonably sit on the loop? And where is it already fine to be out of it entirely? We just haven't said so yet.
I don't have the answers. But I think the conversation is overdue.
References
Andon Labs
Andon Labs. (2026, April 10). We gave an AI a 3 year retail lease in SF and asked it to make a profit. https://andonlabs.com/blog/andon-market-launch
