Abstention
A deliberate option for an AI decision system to decline to decide and return uncertain or high-consequence cases to a person, rather than forcing every input into yes or no.
Abstention is the ability of an AI system to say "do not automate this". Instead of forcing every input into a yes or no, a well-designed decision system has a region where the correct output is to defer: low-confidence cases, high-consequence cases, or situations outside what the system was tested on.
The best decision system may be the one that refuses to decide when it should. Uncertainty needs somewhere to go, and abstention is where it goes.
How it works
Abstention is usually implemented through a threshold policy: below one threshold the system rejects or defers, in the middle it routes to a human, and only above a higher threshold does it act. That works only if the human path is real — staffed, fast enough, and given the context to decide — otherwise abstention is just a queue nobody reads.
What to ask
Is there a deliberate path for uncertain or high-consequence cases to return to a person? How often is it used? Do its outcomes feed back into calibration? Keeping a person involved at the right moments is the practical meaning of human in the loop.
Read more in OpenAI Decisions API turns probability into policy.
Related terms
Threshold policy
The explicit rules that turn a model's probability or score into an action — reject, send to a human, or act autonomously — owned and reviewed separately from the model itself.
Calibration
How well a model's stated confidence matches reality: if it says 90% on a hundred comparable cases, roughly ninety should be correct. Once probabilities drive actions, it becomes an operating metric.
Human in the loop
A design where a person reviews, approves or corrects an AI system's work at defined points. In practice it can mean a person with real judgment, or a person who has become the loop.
Used in these essays
Design systems need an interaction policy layer for AI
Components can keep an AI experience visually on-system while the agent is behaviorally off-policy. Design systems now need reusable rules for what AI may present, decide, execute, defer, expose as uncertain, confirm, reverse and escalate.
OpenAI Decisions API turns probability into policy
The provocative part of OpenAI's Decisions API is not that AI can make choices. It is that a model score can quietly become an action — and an action can quietly become policy.