> For the complete documentation index, see [llms.txt](https://docs.lpp-minduniverse.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.lpp-minduniverse.org/lingua-pactum-protocol-lpp-documentation/1.constitutional-foundation/1.3-human-sovereignty.md).

# 1.3 Human Sovereignty

Human sovereignty is a structural requirement of LPP.

It is not equivalent to requiring a human to manually approve every low-level operation.

It means that the legitimacy of consequential AI-triggered action must ultimately originate from authority that is recognizable within human or institutional governance.

#### Human Authority

An AI system may:

* analyze,
* recommend,
* plan,
* transform information,
* decompose tasks,
* propose actions,
* or participate in execution workflows.

It does not thereby acquire sovereign authority.

The distinction is:

```
AI may generate an Action Candidate.

AI may not create the legitimacy
that makes the Action Candidate admissible.
```

The ability to propose an action is therefore distinct from the authority to authorize that action.

#### Institutional Authority

Human sovereignty does not require all authority to originate from an individual acting personally at execution time.

Legitimate authority may exist through recognized institutional structures.

Examples may include:

* organizations,
* regulated institutions,
* delegated corporate roles,
* government bodies,
* formally authorized operators,
* or other identifiable legal entities.

The constitutional requirement is that authority remains attributable to a legitimate human-recognizable governance source.

Institutional delegation does not convert AI into the sovereign source.

#### Non-Self-Authorizing AI

A central constitutional prohibition is:

> **AI systems must not become self-authorizing.**

An Execution Plane may propose:

```
I intend to execute action E.
```

It may not validly conclude:

```
I consider action E legitimate,
therefore I authorize myself to execute E.
```

Model confidence cannot create authority.

Agent reasoning cannot create authority.

Successful planning cannot create authority.

Runtime access cannot create authority.

Tool possession cannot create authority.

A system may determine that an action is technically possible.

That does not establish that the system has legitimate authority to cause the action.

#### Human Sovereignty as a Traceability Requirement

Human sovereignty must therefore be visible in the governance chain.

At minimum, consequential execution should remain capable of answering:

```
Who or what institution authorized this?

What was actually authorized?

Under what scope?

For how long?

Can the authority be revoked?

Who remains accountable?

Can the legitimacy chain be reconstructed?
```

If these questions cannot be resolved, the governance problem is not merely missing metadata.

The legitimacy basis itself may have failed.