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 decision-making remain in the hands of the user.

• Our solutions provide contextualized analyses and probabilistic indicators designed to deepen 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 oversight by design

Human oversight is integrated by design across all our systems to ensure that people retain control at every stage of the decision-making process.

• Generated analyses remain interpretable and can be reviewed by users at any time.
• Users can interpret, contextualize or disregard system outputs whenever necessary.
• No irreversible or binding decision is made by our systems.

This approach is aligned with the requirements of the European AI Act and with international best practices applicable to high-risk AI systems.

1.3 Fairness, bias awareness and scientific integrity

Detecting, measuring and reducing potential bias, as well as preventing inappropriate use, 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 complete 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 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 retain full strategic, technological and operational autonomy.
• Our solutions can be deployed on sovereign infrastructures, including:
  • cloud environments,
  • private clouds,
  • on-premises deployments, depending on each client’s requirements and constraints.

While 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 Full ownership of our technology stack

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

• We do not rely on proprietary artificial intelligence APIs provided by third parties.
• 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,
• full control over system behavior and evolution over time.

2.3 Research-driven and scientifically documented AI

Our scientific approach is built 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 strengthens Stern Tech’s position as a French deeptech company developing behavioral analysis solutions supported by rigorous, documented research carried out with recognized academic partners.

Governance, expertise and ethical oversight

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.

The board:

• contributes to the development and evolution of our research roadmap,
• participates in reviewing and assessing our methodological directions,
• helps maintain methodological consistency and scientific rigor 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 ensures:

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

3.3 Energy efficiency and privacy by architecture

Ethical artificial intelligence also requires responsible use of resources.

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

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

This architecture makes it possible to:

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