ELNAV.AI d.o.o.
Last updated: 12.12.2024.

At ELNAV.AI, we develop artificial intelligence systems that support critical maritime operations, including speech-based emotion recognition, fatigue detection, and navigational safety. As an innovator in safety-critical environments, we are deeply committed to ethical AI development and the protection of human rights on board.

This policy outlines our principles and safeguards in line with the EU AI Act, Charter of Fundamental Rights of the European Union, and UN Guiding Principles on Business and Human Rights.

1. Human-Centric Design

We design AI systems to augment, not replace, human expertise.

• All AI outputs—such as stress alerts or helm deviation warnings—are advisory only.
• Final decisions remain with human operators (officers, pilots, captains).
• Systems are deployed in a human-in-the-loop model, ensuring oversight, responsibility, and accountability

 

2. Transparency and Explainability

We strive to make our systems understandable, both for users and for external stakeholders.

• All AI models are documented with clear descriptions of how they work, including inputs, outputs, and logic.
• Alerts are accompanied by brief, human-readable explanations (e.g., “Prolonged eye closure detected – possible fatigue”).
• Users receive training to interpret AI alerts correctly and respond accordingly.

 

3. Fairness, Non-Discrimination & Bias Mitigation

We actively monitor and mitigate risks of bias in our AI systems.

• Speech Emotion Recognition models are tested for demographic bias (e.g., language, accent, gender).
• We analyze and adjust training datasets to ensure diverse representation.
• Quarterly reviews are conducted to identify potential false positives or discriminatory outcomes.

 

We do not support use cases that:

• Profile individuals based on gender, ethnicity, or emotional states for performance evaluation.
• Use emotion recognition to make automated decisions affecting crew rights or employment.

 

4. Safety and Risk Management

We apply our values inside the company as well:

• Digital-first, paper-light operations
• Remote/hybrid work to reduce commuting emissions
• Preference for sustainable suppliers and EU-based production
• Continuous training on sustainability and ethics for our team

 

5. Accountability and Governance

• Ethical oversight is the responsibility of the CEO and a designated Compliance Lead.
• We follow internal protocols for incident reporting, user feedback, and ethics evaluations in EU-funded projects.
• Partners and clients are required to uphold equivalent standards when integrating our technologies.

 

6. Privacy and Human Dignity

Although addressed in detail in our [Data Protection & Privacy Policy], we reaffirm here:

• We do not record or store raw biometric data (e.g., video, voice) unless explicitly required for research with informed consent.
• We prioritize privacy-preserving AI methods.
• Our systems are designed to support well-being, not to surveil or penalize.

 

7. Continuous Improvement

ELNAV.AI’s solutions are deployed in safety-critical contexts (e.g., ship navigation, port approach).

• All systems undergo rigorous testing and simulation before deployment.
• Fail-safes are in place to prevent over-reliance on AI (e.g., timeout-based escalation if no human acknowledges an alert).
• Systems are continuously monitored and improved based on user feedback and technical audits.