The adoption of AI in healthcare is not determined solely by the capabilities of AI models. It also depends on the ability to integrate these technologies safely and effectively into everyday clinical workflows.
Ambiti8n aims to transform research outcomes into practical solutions that support healthcare professionals in diagnostic activities while ensuring system security, transparency and meaningful human oversight throughout the decision-making process.
In such a complex and highly regulated sector, progress creates value only when it is accompanied by trust. For this reason, the project places responsibility and the protection of people at its core, demonstrating how federated AI, sovereign infrastructure and modern data management methodologies can contribute to the development of dedicated, reliable and sustainable digital solutions across Europe.
Villanova.ai's contribution to Ambiti8n
Within Ambiti8n, our team developed the project's AI language technologies together with the governance mechanisms that ensure AI is used safely, transparently and in compliance with applicable regulations.
Our work is centered on trustworthiness, based on the understanding that confidence in AI systems depends first and foremost on the robustness and transparency of the processes behind them.
At the heart of our contribution are:
- Villanova 2B — a sovereign European large language model optimized for clinical reporting.
- Altea — our AI Governance framework, which embeds compliance, explainability and risk controls directly into the platform's architecture, ensuring alignment with the GDPR and the European AI Act.
From AI to better clinical practice
What do clinicians actually need? They certainly do not need more raw data. What they need instead is reliable, interpretable and actionable information that supports clinical assessment and decision-making.
In the Ambiti8n use case, the AI model identifies potential sleep apnea events, while Villanova 2B transforms these findings into structured clinical reports using terminology, language and formats specifically designed for healthcare professionals.
AI is designed to assist, not replace, clinical expertise.
For this reason, the system follows a human-in-the-loop approach: technology enhances physicians' decision-making capabilities while preserving their expertise, autonomy and ultimate responsibility.
🎥 In the following video, Paolo Budroni explains how Villanova 2B and Altea fit into the federated Ambiti8n ecosystem.
Compliance by design: embedding governance into AI architecture
The European AI Act, GDPR, risk management and human oversight are not afterthoughts — they are the principles that guide the system architecture from the very beginning.
This is particularly important in high-risk domains such as healthcare, where regulatory compliance must be built into the entire development lifecycle. AI systems must satisfy stringent requirements from the outset, including:
- Data quality and traceability
- Transparency and explainability of AI outputs
- Human oversight
- Risk management
- Continuous monitoring and comprehensive documentation
This is why Ambiti8n adopts a compliance-by-design approach, creating a single integrated ecosystem rather than a collection of disconnected components.
A European ecosystem for trustworthy AI
Building sovereign infrastructure, federated AI, governance frameworks, orchestration and clinical workflows requires complementary expertise. This is made possible through the collaboration of the eight teams involved in the project.
The challenge is not simply to develop increasingly powerful AI models, but to build AI systems that can be responsibly integrated into real-world environments such as healthcare.
The outcome achieved goes beyond research: it represents a concrete model for the adoption of secure, trustworthy and sovereign European AI, ready to support healthcare applications across Europe.