human-in-the-loop
In plain wordsDescribes AI systems in which a person checks, corrects or approves the machine's work at key points, rather than letting it act entirely on its own.
- 1
Oversight and automation
Describing a system in which a human must review, approve or be able to override automated outputs before they take effect, as when a doctor confirms an AI-flagged diagnosis or an agent asks permission before running a command. Writers on autonomy distinguish human-in-the-loop (a human authorizes each action), human-on-the-loop (the system acts on its own while a human monitors and can intervene) and human-out-of-the-loop (full autonomy); the distinction is central to debates over autonomous weapons. Effective oversight requires that the human have time, information, competence and authority to disagree; otherwise automation biasautomation bias, n.: The habit of trusting a computer's suggestion too much: following it even when it is wrong, or missing problems because the machine did not flag them. reduces review to a rubber stamp. The EU AI ActEU AI Act, prop. n.: The European Union's law on artificial intelligence, the first broad AI law anywhere: it bans some uses, sets strict checks for risky ones and puts duties on the makers of the biggest AI models. requires human oversighthuman oversight, n.: Keeping people able to watch, understand, question and, if necessary, stop or overrule an AI system, rather than letting it act unchecked. measures for high-risk systemshigh-risk AI system, n.: Under EU law, an AI system used where mistakes could seriously hurt people's safety or rights — in hiring, schooling, loans, policing, borders or products like medical devices — and which must therefore pass strict checks..
Payments over $10,000 stay human-in-the-loop: the agent drafts them, but a person approves each one.
- 2
Machine learning
Describing training and data workflows that put human judgment inside the learning process: labelingdata labeling, n.: The work of attaching correct answers to examples — often done by large paid workforces, sometimes by experts — so that AI systems can learn from them or be tested against them. data, correcting model outputs, choosing which examples to annotate (active learningactive learning, n.: A way to train a model with fewer labeled examples by letting it choose which unlabeled examples a person should label next, usually the ones it is most unsure about.) or ranking responses for reinforcement learning from human feedbackreinforcement learning from human feedback, n.: A way of training AI using people's judgments: humans compare pairs of answers, a scoring model learns their preferences, and the AI is then trained to produce answers that score well..
Origin
From control engineering and military simulation, where 'man-in-the-loop' described systems and simulations with a human operator inside the feedback loop. The gender-neutral form spread later, and the in/on/out-of-the-loop taxonomy became prominent in debates on lethal autonomous weapons around 2012.
Usage note
A human 'in the loop' does not guarantee meaningful control: speed, volume and opaque outputs can make review nominal. Not to be confused with reinforcement learning from human feedbackreinforcement learning from human feedback, n.: A way of training AI using people's judgments: humans compare pairs of answers, a scoring model learns their preferences, and the AI is then trained to produce answers that score well., which uses human judgments during training rather than operation.
Related
- human oversighthuman oversight, n.: Keeping people able to watch, understand, question and, if necessary, stop or overrule an AI system, rather than letting it act unchecked.
- automation biasautomation bias, n.: The habit of trusting a computer's suggestion too much: following it even when it is wrong, or missing problems because the machine did not flag them.
- active learningactive learning, n.: A way to train a model with fewer labeled examples by letting it choose which unlabeled examples a person should label next, usually the ones it is most unsure about.
- levels of autonomylevels of autonomy, n.: Ladders that grade how independently an AI system acts — from a tool a person fully controls, through suggesting and asking for approval, up to acting on its own while people merely watch.
- permission systempermission system, n.: The rules that decide what an AI agent may do on its own, what it must ask a person about first, and what it may never do — like a phone's settings for which apps can use the camera.
- reinforcement learning from human feedbackreinforcement learning from human feedback, n.: A way of training AI using people's judgments: humans compare pairs of answers, a scoring model learns their preferences, and the AI is then trained to produce answers that score well.
- centaurcentaur, n.: A human and an AI working as a team, each doing what it does best, like a chess player who consults a computer before every move. The name comes from the mythical half-human, half-horse creature.
- autonomyautonomy, n.: How much an AI system can act on its own without a person approving each step, from merely making suggestions to carrying out long tasks with no supervision at all.
Mentioned in
copilotcopilot, n.: An AI assistant built into software that works alongside a person — suggesting code, drafting text or answering questions — while the person stays in charge, like a co-pilot beside the captain., lethal autonomous weapons systemlethal autonomous weapons system, n.: A weapon that, after a person sets it going, picks out and attacks targets — including people — on its own, with no human deciding each strike. Critics call them killer robots., active learningactive learning, n.: A way to train a model with fewer labeled examples by letting it choose which unlabeled examples a person should label next, usually the ones it is most unsure about., automated decision-makingautomated decision-making, n.: When a computer system, rather than a person, makes or effectively makes a decision about someone — such as approving credit, screening job applicants or flagging possible welfare fraud., automation biasautomation bias, n.: The habit of trusting a computer's suggestion too much: following it even when it is wrong, or missing problems because the machine did not flag them., autonomyautonomy, n.: How much an AI system can act on its own without a person approving each step, from merely making suggestions to carrying out long tasks with no supervision at all., centaurcentaur, n.: A human and an AI working as a team, each doing what it does best, like a chess player who consults a computer before every move. The name comes from the mythical half-human, half-horse creature., data labelingdata labeling, n.: The work of attaching correct answers to examples — often done by large paid workforces, sometimes by experts — so that AI systems can learn from them or be tested against them., false negativefalse negative, n.: A miss: a test or AI system gives the all-clear, for example no illness, no fraud or no danger, when the problem is in fact there., human oversighthuman oversight, n.: Keeping people able to watch, understand, question and, if necessary, stop or overrule an AI system, rather than letting it act unchecked., human-centered AIhuman-centered AI, n.: An approach to AI that puts people's needs, abilities and control at the center: systems designed to help humans do things better rather than to replace or overrule them., international humanitarian lawinternational humanitarian law, n.: The rules of international law that limit how wars may be fought, protecting civilians, the wounded and prisoners. They apply to every weapon, including ones that use AI., levels of autonomylevels of autonomy, n.: Ladders that grade how independently an AI system acts — from a tool a person fully controls, through suggesting and asking for approval, up to acting on its own while people merely watch., permission systempermission system, n.: The rules that decide what an AI agent may do on its own, what it must ask a person about first, and what it may never do — like a phone's settings for which apps can use the camera., teleoperationteleoperation, n.: Controlling a robot or machine from a distance, with a person steering it in real time using a joystick, a matching 'puppet' arm, a VR headset or a screen., algorithm aversionalgorithm aversion, n.: Losing faith in a computer's predictions after watching it slip up once or twice, and turning back to human judgment even though the machine remains the better forecaster., co-intelligenceco-intelligence, n.: Ethan Mollick's term for working and thinking alongside AI as a partner — treating it like a capable but odd colleague whose strengths and blind spots you learn through experience., concept bottleneck modelconcept bottleneck model, n.: A model made to first predict human-understandable concepts — such as 'bone spur present' on an X-ray — and then decide only from those concepts, so people can inspect and correct them., data flywheeldata flywheel, n.: A cycle in which a product's users generate data that improves its AI, the better AI attracts more users, and those users generate still more data., LangGraphLangGraph, prop. n.: A free software library for building AI agents as flowcharts of steps that share a memory, so developers control exactly how an agent moves between steps, pauses and resumes., moral crumple zonemoral crumple zone, n.: When a largely automated system goes wrong and the human who was supposed to supervise it gets blamed, even though they had little real say; named after the car part that soaks up a crash.