An early draft we are still building, shared to start a conversation rather than settle it. If the examples or the wording do not match your experience, tell us at hello@hunteraicommons.com and help us make it reflect the community better.
Your home wifi drops out. You ring your telco, a chatbot answers, tells you plainly that it is a bot, and walks you through a router reset. That might not bother you at all, as long as it is quick and helps you get to the fix you need.
Now picture someone ringing a mental health crisis line at two in the morning. Here, a lot of us would expect a person on the other end, and a bot might feel like a let-down at exactly the wrong moment.
Two phone calls, and two very different expectations. Most situations sit somewhere in between, and someone has to decide where the line falls.
It is a phrase you will hear more and more. "Human in the loop" just means a real person stays involved in a decision or a conversation, rather than it being handed over entirely to a machine.
Sometimes we want a person to be the one who decides, who answers, or who can be asked "why". Sometimes a machine is perfectly fine, and a person in the way would only slow things down. The tricky part is that it changes from one situation to the next, and we rarely stop to say where our own line sits.
So here is that line. It runs from things we are happy to let a machine handle, on the left, to things where we expect a person every time, on the right. Scroll through, and see where each one lands.
The easy end
The router reset. The bot answers, sorts it out, you hang up. For a job like this, a lot of people would not mind a bot at all. Hold this spot in mind. It is the easy end.
A small step along
You ask which week your recycling goes out, and a bot answers in seconds. Plenty of us would be happy with that. It sits near the easy end too.
A little further
Booking through an app is handy. But you might feel a small tug to the right. It touches your health, and sometimes the thing you need to say does not fit in the form.
It moves
Your home loan is knocked back and you want to know why, but all that answers is an automated message. For a lot of people, the bot starts to feel wrong here. The stakes have climbed, and the line moves with them.
Further still
A letter says you owe money and you are sure it is wrong, but the only thing answering is an algorithm. We have seen this go badly before. Many people would expect a person here, one who can listen and be accountable.
Near the far end
A test comes back with something serious on it. Not many of us would want that news from a machine. Not because a machine could not say the words, but because the moment asks for a person to be there.
And here
A teenager, awake late, tells a chat companion they feel worthless. This sits close to the far end. A lot of us would want a child in that moment to be able to reach a human who can help.
The far end
Here the point is not really to solve a problem. It is to be heard by another person. This is the far end of the line, where a lot of us feel a human should always be the one who answers.
Go back to the easy end. Part of why the telco chatbot felt fine is that it was honest about what it was. You knew. You could ask for a person if you wanted one.
Now change one detail. Imagine the bot had introduced itself as a person called Sarah, and you only found out later. Even for a router reset, a lot of us would feel something had been taken. Being misled is its own kind of harm, however small.
So there is a second expectation, running underneath the first. People expect to know whether they are dealing with a person or a machine. Being told is not a courtesy. It is part of the deal.
Every day, businesses and governments decide where the line goes. A bank automates a refusal. A service centre hands the after-hours queue to a bot. A platform leaves a distressed teenager with a chatbot and no clear way through to a person. Sometimes that is fine. Sometimes it is done to cut costs, where a human was the whole point.
These expectations are real, and they are widely shared. The work is to name them clearly, and to hold the governments and businesses making these choices to them. A community is allowed to decide where its lines sit, and to expect to be heard.
It is built by people and shaped by choices, so we can have a say. Where a human stays in the loop is one of those choices, and it is not only for someone else to settle.
That is what the Commons is for. Working out, together, where our lines sit, and helping make sure the systems shaping our lives are held to them.
Shaping AI, together, for the common good.