Behavioral AI designed to remain human.

Responsible use of artificial intelligence

1.1 Decision support only

All Stern Tech solutions are designed as decision-support systems, ensuring that responsibility and final decisions remain in the hands of the user.

• Our solutions provide contextualized analyses and probabilistic indicators designed to enrich the understanding of situations and support human decision-making.
• They do not make autonomous or automated decisions.
• They do not replace human judgment, expertise or user responsibility.

Final decisions always remain under human responsibility, in accordance with applicable legal requirements and ethical principles.

1.2 Human supervision by design

Human supervision is embedded by design across all our systems, ensuring that people retain control at every stage of the decision-making process.

• The analyses generated are interpretable and can be reviewed at any time by users.
• Users can interpret, contextualize or dismiss the results provided by the system at any time.
• No irreversible or binding decision is made by our systems.

This approach is aligned with the requirements of the European Artificial Intelligence Act, as well as with international best practices applicable to high-risk AI systems.

1.3 Fairness, bias awareness and scientific integrity

Detecting, measuring and reducing potential biases, as well as preventing inappropriate uses, are integral parts of our AI governance approach.

• Our models are developed according to strict dataset-governance principles.
• Their performance and limitations are continuously monitored.
• Residual risks are acknowledged transparently: no AI system can guarantee the total absence of bias or error.

Our role is to provide reliable decision-support tools, not to dictate outcomes.

Technological sovereignty and strategic independence

2.1 A fully sovereign AI technology

Stern Tech is a fully independent French company, 100% owned by French shareholders.

• We are not controlled by any foreign group or technology conglomerate.
• We maintain complete strategic, technological and operational autonomy.
• Our solutions can be deployed on sovereign infrastructures, including:
  • cloud environments,
  • private clouds,
  • on-premise deployments, depending on each client’s requirements and constraints.

Although our services may be hosted on platforms such as Microsoft Azure within the European Union, our core technology does not depend on any proprietary cloud-provider API.

2.2 Exclusive ownership of our technological core

Stern Tech exclusively owns its software, algorithms, models and proprietary technologies.

• We do not depend on any proprietary artificial-intelligence API provided by a third party.
• When open-source components are used, they are:
  • carefully selected,
  • adapted, enriched or extended internally,
  • integrated in strict compliance with Apache 2.0 and MIT licenses.

This approach ensures:

• long-term technological independence,
• a high level of transparency, traceability and auditability,
• complete control over the behavior of our systems and their evolution over time.

2.3 AI grounded in research and scientifically documented

Our scientific approach is based on long-term research collaborations with leading French academic institutions, including:

These collaborations allow us to develop AI models grounded in peer-reviewed scientific research, rigorous mathematical frameworks and experimental validation.

This approach reinforces Stern Tech’s position as a French deeptech company developing behavioral-analysis solutions supported by demanding, documented research and built with recognized academic partners.

Governance, expertise and ethical supervision

3.1 A dedicated scientific advisory board

Our scientific advisory board brings together leading researchers and experts in behavioral sciences, applied mathematics and cognitive modeling to support the development and validation of our technologies.

This board:

• contributes to the construction and evolution of our research roadmap,
• participates in the review and evaluation of our methodological orientations,
• helps ensure the methodological coherence and scientific rigor of our work over time.

3.2 Designed and developed in France

Based in France, Stern Tech develops technologies that are:

• designed,
• developed,
• tested,
• and governed entirely in France.

This approach helps ensure:

• compliance with applicable European regulations,
• alignment with French and European ethical principles,
• reinforced protection against extraterritorial dependencies.

3.3 Energy efficiency and privacy protection by architecture

Ethical artificial intelligence also requires responsible use of resources.

Our systems are designed to limit their environmental impact while ensuring responsible use of data:

• Behavioral analysis is primarily performed locally on user devices, including smartphones, tablets, computers or embedded vehicle systems.
• Processing is performed in volatile memory, without persistent local storage of data.
• Only aggregated and non-identifying results are transmitted when necessary.
• Unnecessary data is automatically removed from memory through a proprietary mathematical model developed in collaboration with French public research.

This architecture makes it possible to:

• significantly reduce energy consumption,
• limit data circulation and transfer,
• strengthen privacy protection by design.