Why Intention Matters for Judgement

Courts do not judge actions alone. They judge the intention behind them. Systems, however, act without preserving that intention.

This is not philosophy. It is how law works.

In criminal law, a person is not judged by what happened alone. The court has to establish both the act and the mental state behind it. This principle — mens rea, the "guilty mind" — is foundational: an act without intention is not judged the same way as an act with intention, because actions don't explain themselves.

The Origin of Responsibility

Intention is the point where responsibility begins. In law, intention is defined as the decision to bring about a specific consequence — and that definition reveals something critical: responsibility isn't assigned at the moment of outcome. It's traced back to the moment of intent. Without that trace, judgement collapses.

Why the Act Is Not Enough

Two outcomes can be identical — an accident, or a deliberate act. The result may be the same. The judgement is not, because judgement was never about what happened. It's about what was meant to happen.

Now Look at Systems

Modern systems produce outcomes. AI systems generate actions, decisions, summaries, and recommendations. They appear to act. But they do not originate intention — they receive it, transform it, and propagate it.

The Structural Problem

At every step in a system, intention is interpreted, context is reduced, and meaning is reconstructed. This isn't failure. It's how systems work. But it creates a real consequence:

The Point of Fragmentation

Systems don't fail because they act. They fail when they lose track of what they were supposed to act for. The moment the originating intention is no longer accessible, meaning becomes unstable, outputs become disconnected, and responsibility becomes unclear. That's the real fragmentation point.

From Law to System Design

Law enforces a simple rule: no judgement without intention. Systems do not — they produce outcomes without preserving the intention that made those outcomes meaningful in the first place. This is why alignment is unstable. Not because systems are uncontrollable, but because intention does not survive the system.

The Requirement

If systems are to support judgement, responsibility, and alignment, they need to adopt the same principle as law: at every step, a system must carry forward the originating intention, revalidate it, or stop the process. Otherwise, the outcome is no longer anchored in its origin.

The Real Alignment Problem

The fear is that AI will develop its own intention. The reality is simpler:

And when that happens, we can no longer judge the outcome properly.

Because that's what judgement depends on. Without it, neither responsibility nor alignment can hold.