MoodMe EU AI Act Compliance
Table of Contents
MoodMe Belgium SRL ("MoodMe", "we", "us", or "our") develops Emotion AI software that analyzes facial expressions to generate emotional insights. We are committed to responsible AI development and compliance with the European Union Artificial Intelligence Act (Regulation (EU) 2024/1689) (the "EU AI Act").
This statement describes how MoodMe aligns its products and practices with the requirements of the EU AI Act.
1. Overview
The EU AI Act establishes a harmonized legal framework for the development, deployment, and use of artificial intelligence systems across the European Union. It introduces a risk-based classification system and imposes obligations on providers and deployers of AI systems proportionate to the level of risk.
MoodMe proactively aligns its Emotion AI technologies with the EU AI Act to ensure our customers can deploy our solutions with confidence across regulated markets.
2. Scope & Applicability
This statement applies to all MoodMe AI-powered products and services, including:
- Emotion recognition SDKs and APIs
- Face tracking and analysis modules
- MoodMirror and related applications
- Any derivative or embedded AI components provided to customers
Role Determination
Under the EU AI Act, MoodMe may act as:
| Role | When |
|---|---|
| Provider | When MoodMe develops and places an AI system on the market or puts it into service under its own name or trademark |
| Supplier of components | When MoodMe supplies SDKs or modules integrated by customers into their own AI systems |
The specific obligations applicable depend on the role and the risk classification of the resulting system.
3. Risk Classification
The EU AI Act categorizes AI systems into four risk tiers:
| Risk Level | Description | MoodMe Relevance |
|---|---|---|
| Unacceptable | Prohibited AI practices | Not applicable — MoodMe does not engage in prohibited practices |
| High | AI systems in critical domains (e.g., biometrics, law enforcement) | May apply when emotion recognition is used in regulated contexts |
| Limited | AI systems with transparency obligations | Applicable — emotion recognition systems require transparency disclosures |
| Minimal | Low-risk AI applications | May apply for general entertainment or wellness use cases |
Emotion recognition systems are explicitly referenced in the EU AI Act. When deployed in workplace or educational settings, they are classified as high-risk and subject to enhanced obligations (Article 6, Annex III). In all contexts, they require transparency notifications to individuals (Article 50).
4. Privacy & Safety by Design
MoodMe's architecture embodies privacy-by-design and safety-by-design principles:
- On-device processing: Emotion analysis runs locally on the user's device — facial data is not transmitted to MoodMe servers
- No biometric storage: Face images and biometric identifiers are processed in real time and discarded
- Data minimization: Only aggregated, non-identifiable outputs are generated by default
- User control: End users and customers retain control over what data, if any, is stored or transmitted
This architecture reduces the risk profile of MoodMe's technology and supports compliance with both the EU AI Act and the GDPR.
5. Transparency & Documentation
MoodMe is committed to transparency in its AI systems, in accordance with Article 50 of the EU AI Act:
Notification to Users
MoodMe provides clear mechanisms and guidance so that deployers can inform individuals when they are interacting with an emotion recognition system, as required by Article 50(3).
Technical Documentation
MoodMe maintains comprehensive technical documentation for its AI systems, including:
- System description and intended purpose
- Architecture and processing logic
- Data requirements and handling practices
- Performance metrics and known limitations
- Instructions for safe deployment
Record-Keeping
Where applicable, MoodMe implements logging capabilities that enable traceability and auditability of AI system operations, proportionate to the risk level and deployment context.
6. Data Governance
In line with Article 10 of the EU AI Act, MoodMe applies rigorous data governance practices:
- Training data quality: Models are developed using curated, diverse, and representative datasets
- Bias monitoring: Ongoing evaluation for demographic and contextual biases
- Data documentation: Training data sources, characteristics, and processing methods are documented
- Purpose limitation: Data used for model development is not repurposed without appropriate safeguards
7. Human Oversight
MoodMe designs its AI systems to support meaningful human oversight, as required by Article 14:
- AI outputs are designed to assist — not replace — human decision-making
- Systems include configurable thresholds, confidence indicators, and interpretability features
- Deployers are provided with guidance on appropriate human-in-the-loop practices
- Emergency stop and override mechanisms are supported where applicable
8. Accuracy, Robustness & Cybersecurity
In compliance with Article 15, MoodMe ensures:
Accuracy
- Models are validated against diverse benchmarks and real-world conditions
- Performance metrics and known limitations are disclosed in technical documentation
- Continuous improvement through feedback loops and model updates
Robustness
- Systems are tested for resilience against adversarial inputs and edge cases
- Graceful degradation is implemented for low-quality or ambiguous input data
- Fallback behaviors are defined and documented
Cybersecurity
- On-device processing architecture minimizes attack surface
- SDK integrity is protected through code signing and secure distribution
- Regular security assessments and vulnerability management processes are in place
9. Prohibited Practices
MoodMe does not engage in any AI practices prohibited under Article 5 of the EU AI Act, including:
- Subliminal manipulation or deceptive techniques
- Exploitation of vulnerabilities of specific groups
- Social scoring by public authorities
- Real-time remote biometric identification in publicly accessible spaces for law enforcement (except as permitted by law)
- Untargeted scraping of facial images
- Emotion recognition in workplaces or educational institutions without lawful basis and appropriate safeguards
MoodMe's technology is designed for voluntary, informed use. We actively prohibit use of our Services in contexts that would constitute prohibited AI practices under the EU AI Act.
10. Conformity Assessment & Standards
MoodMe is actively preparing for conformity assessment requirements under the EU AI Act:
- Quality management system: Internal processes aligned with EU AI Act Article 17 requirements
- Risk management: Continuous risk identification, analysis, and mitigation throughout the AI system lifecycle (Article 9)
- Harmonized standards: Monitoring and adopting emerging CEN/CENELEC standards for AI systems as they become available
- Post-market monitoring: Systematic processes to monitor performance, collect feedback, and address issues after deployment (Article 72)
As the EU AI Act implementation timeline progresses (2025–2027), MoodMe will continue to update its compliance measures in line with published standards and regulatory guidance.
11. Updates to This Statement
This compliance statement may be updated to reflect changes in the EU AI Act's implementing measures, technical standards, or MoodMe's products and practices. The "Last Updated" date reflects the current version.
12. Contact
For questions regarding MoodMe's EU AI Act compliance:
Email: support@mood-me.com
Contact Page: www.mood-me.com/contact-us