Term · The values layer
The values layer for AI
Every model applies values. The values layer is the part of the system where those values are written down where you can see them, argue with them, and change them, instead of leaving them buried where you can only discover them by being harmed.
The values layer for AI is the part of a system where what it should care about is made explicit and contestable. Today, in most systems, values are implicit: smuggled in through training data, a scattered policy, and a hundred small choices nobody recorded. The values layer is the proposal to pull those choices into the open, into a place you can point at, so that the question "why did it decide that" has an answer you can read rather than reverse-engineer.
The layer of an AI system where the values it applies are made explicit, visible, and contestable, rather than left implicit in the weights or scattered across undocumented choices.
Why it needs to be a layer
A value you can point at is a value you can argue with. A value baked into a model is one you can only find by tripping over it. Making values a distinct, legible layer is what turns "trust us" into "here is what it holds, tell us where we are wrong." That is the difference between a system that asks for faith and one that invites scrutiny.
What it is not
It is not a content filter bolted on at the end, and it is not a marketing artifact. It is not someone's private taste imposed quietly. The whole point is the opposite of quiet: values that are surfaced, owned, and open to challenge by the people they affect.
Read on
For the argument behind it, see whose values should AI hold. It sits alongside two siblings: the expertise layer and the trust layer.