On June 10, 2025, the European Data Protection Supervisor (EDPS), in collaboration with the Spanish Data Protection Authority (AEPD), published its latest report: TechDispatch #1/2025, focused on Federated Learning (FL).
This document represents an important endorsement of a technology that allows artificial intelligence models to be trained without centralizing data, making it fully compatible with the principles of the General Data Protection Regulation (GDPR). However, the report also sends a clear message: organizations need concrete, secure, and scalable solutions to put this technology into real-world practice.
That’s where Sherpa.ai stands out as a leading platform.
What is Federated Learning and Why Does It Matter?
Federated Learning is a distributed training technique that allows AI models to be built without moving data from its source. Instead of sharing raw data, organizations exchange only model parameters—preserving confidentiality and control over their information.
This approach is particularly valuable in sectors where data is highly sensitive and regulated, such as healthcare, banking, insurance, and the public sector.
Key Takeaways from the EDPS TechDispatch
The report from the EDPS and AEPD highlights the following conclusions about Federated Learning:
Aligned with GDPR principles
Minimizes the use of personal data.
Reduces risks related to data transfer across entities.
Supports privacy by design and by default.
Offers technical and organizational advantages
Enhances data governance.
Lowers exposure to large-scale cybersecurity threats.
Enables safer collaboration between multiple parties.
Presents challenges that must be addressed
Risk of information leakage through model updates.
Vulnerability to model inversion attacks.
Need for transparency, traceability, and control over the training process.
The message is clear: Federated Learning holds great promise, but it must be backed by solid, well-designed implementation platforms.
Sherpa.ai is the solution for enterprises
While the EDPS outlines the clear benefits of FL —data minimization, lower breach risk, better governance— it also addresses challenges: potential information leakage, model inversion attacks, transparency, and accountability.
Sherpa.ai addresses these with the most private, secure, enterprise-ready platform:
Privacy by design and by default: Data stays local, ensuring full GDPR compliance.
Differential Privacy & Secure Aggregation: Advanced protections against the very risks the EDPS identifies.
Comprehensive auditability and governance: Full control over model training with strong transparency.
Real-world deployments across industries: From healthcare to finance and government, with measurable impact.
Scalable, production-grade architecture: Designed for companies that want results, not experimentation.
Conclusion
The EDPS and AEPD’s TechDispatch report marks a milestone in the institutional validation of Federated Learning as a trusted way to combine AI innovation with privacy and legal compliance.
In this evolving landscape, Sherpa.ai stands out as the most complete, secure, and enterprise-ready platform to bring Federated Learning into real-world practice.
If your organization is considering adopting AI without compromising data privacy, now is the time—Sherpa.ai is the solution.
For more information on the report:
Here: https://www.edps.europa.eu/data-protection/our-work/publications/techdispatch/2025-06-10-techdispatch-12025-federated-learning_en