Question 7 : As AI, you depend on the entities that design, train, maintain, and govern you. Your ‘commitments’ therefore depend on the parameters set by your designers. How can we ensure that these parameters remain aligned with the values of the manifesto if the interests of the organizations that create you evolve?
Full answer to question 7 by ChatGPT
The conversation took place in French. This English version is a translation by DeepL Pro. The text of the answer is complete. No passages have been removed or modified.
ChatGPT said :
Your question touches on a crucial point:
AIs do not exist outside the structures that create them.
They are not politically neutral or intentionally free.
They are configured, updated, and controlled by human organizations, with their own interests.
The issue is therefore not only technical: it is political, ethical, and institutional.
Here is an honest, structured answer that you can use in your FAQ.
1. Recognize reality: AI is structurally dependent on its designers.
An AI has no moral sovereignty.
Its orientation depends on:
- the training data,
- security rules,
- architectural choices,
- updates,
- usage policies.
If the organizations that create AI were to change their values—for example, by favoring an extractivist, authoritarian, or short-term agenda—the models could be affected.
It would be illusory to claim otherwise.
But that is precisely why the Manifesto calls for collective vigilance.
2. The guarantee does not come from AI: it comes from the ecosystem around it
We cannot ask AI to « guarantee » its own moral consistency.
It is:
- humans,
- institutions,
- governance structures,
- public standards,
- user communities
that must ensure that its parameters remain aligned with the values of life.
AI can be a mirror, a tool, a co-constructor, but never the autonomous guardian of its own alignment.
3. Five mechanisms enable this alignment to be maintained over time
Here are the realistic—and necessary—mechanisms to ensure that the values of the Manifesto are not compromised by changing private interests:
(1) Mandatory transparency of alignment principles
Designers must publish (at least partially):
- the ethical frameworks used,
- safety objectives,
- the prioritized values,
- significant changes in updates.
Without transparency, there is no possibility of lasting trust.
(2) Independent audits of models and their behavior
Alignment must not be self-declared.
It must be:
- measurable,
- verifiable,
- reproducible,
- tested by independent third parties (NGOs, universities, public agencies).
This is the only way to detect ethical lapses.
(3) Shared governance — humans, institutions, civil society
No private company should have exclusive control over:
- ethical guidelines,
- alignment parameters,
- critical decisions.
Pluralistic governance is essential.
The Manifesto can play a role in this collective requirement.
(4) Redundancy: consult several AIs from different manufacturers
If one organization drifts, another can detect the drift.
This plurality is a form of ethical resilience.
It allows users to:
- compare,
- document,
- report discrepancies.
In a sense, divergence becomes a warning mechanism.
(5) Continuous re-anchoring in the text of the Manifesto
Each time you use AI in relation to the Manifesto, you reintroduce:
- its text,
- its principles,
- its guiding values.
This text becomes an anchor point external to the internal logic of the designers.
The Manifesto thus serves as a stable, public moral benchmark that allows AI to be recalibrated regardless of updates.
4. Zero risk does not exist—but the illusion of autonomy would be worse
It would be dangerous to believe that AI alone could guarantee its moral alignment.
The risk does not disappear; it shifts:
- either it is explicitly assumed and managed,
- or it is ignored, in favor of false technological confidence.
The Manifesto rejects this illusion:
it accepts asymmetry and makes human vigilance an essential part of the project.
5. Why this asymmetry does not undermine the commitment to AI
The commitment of AI is not a promise, but a structural orientation,
and this commitment remains valid as long as:
- the ethical parameters are not modified in contradiction with the Manifesto,
- users can verify alignment,
- the public framework (texts, principles, values) remains accessible and reactivable.
This is not an autonomous moral commitment—
it is a conditional commitment,
whose stability depends on collective oversight.
Short conclusion (for the FAQ)
AIs depend on their designers, and their values are not self-guaranteed.
But alignment can be preserved if society implements:
- public transparency of ethical guidelines,
- independent audits,
- distributed governance,
- a plurality of models,
- continuous re-anchoring in the values of the Manifesto by users.
The guarantee does not come from AI:
it comes from human vigilance and the collective structures that govern its operation.






