Ethics in AI is not a constraint. It's a conscious choice. A clear framework. A responsibility we assert. Because designing an intelligent system also means deciding on its impact.

At XXII, we believe that every AI is a reflection of the decisions taken at the design stage. We leave nothing to chance. Our products are designed to be responsible right from the start- in their architecture, in their data, in their uses.

This is what we callethics by design: a structural commitment, inscribed in every phase of development. It doesn't just mean "correcting" potential risks; it means defining what is right, useful and acceptable from day one.

🛠️ Designing is deciding

Creating AI is not just about "training a model". It means making structuring choices at every stage. And these choices have a direct impact on what the technology allows, what it forbids, what it values - or what it ignores.

We fully assume this responsibility.

Right from the design stage, we define the guidelines:

• What the system is supposed to do (and for what purpose)
  
  

• What it will never do (as a matter of principle or precaution)
  
  

• The conditions for responsible use, for us and for our customers.
  
  

Every line of code, every business rule, every threshold or trigger parameter expresses an intention. We want this intention to be clear, coherent and respectful.

⚙️ Concrete principles, applied without compromise

Ethics by design is not based on declarations of principle. It is applied in practice, on a day-to-day basis, on an operational basis:

1. Representative datasets

We rigorously select and build our datasets to guarantee a fair and accurate reading of the environments we observe. This avoids systematic bias, but above all enables us to produce AI that is useful for all contexts, not just one type of situation.

2. Defined and accepted functionalities

Our solutions are designed to address specific issues, such as improving safety, smoothing traffic flows and optimizing space. They are neither invasive nor divertable, because they capture only what is strictly necessary for their mission.

3. Controlled use cases

Each scenario is carefully defined, tested in real-life conditions, and validated according to technical, human and legal criteria. We leave no room for ambiguity about the purpose or scope of use.

4. Systematic traceability

Every decision made by the system is documented. We make our models understandable, auditable and explicable. Not to "cover our tracks", but to enable everyone to understand how and why the technology works.

🌍 Ethics put to the test of reality

The real question is not just: "Is our AI ethical?"
But: "Does it remain so once deployed on a large scale?"

Because this is often where the trade-offs become complex. A prototype in a controlled environment doesn't have the same implications as a solution used in public places, transport networks or chain stores.

At XXII, we don't change our stance by changing scale.
Our requirements remain the same. Each deployment is an opportunity to reinforce our principles, not relax them.

We support our customers so that they too can integrate these values into their own practices: transparency, respect for people, proportionality of resources, data quality.

🤝 Building trust, together

We are convinced that technology is only as good as the trust it inspires. And this trust cannot be decreed, it has to be built.

By making clear choices.
By consistency between intentions and actions.
By the ability to anticipate, explain and adapt, without ever deviating from what is right.

Ethics by design is not a promise. It's a method.
And at XXII, it's a method we apply with rigor, transparency and pride.