MAIA: The new push for distributed, sustainable and secure artificial intelligence in the productive and public sectors

• The project is driven by i2CAT and Blueroom Innovation, with the support of ACCIÓ, as part of the Projectes R&D Empresarial Programme.

In a context where adopting artificial intelligence (AI) often requires sending critical data to external cloud servers, security, privacy, and high energy consumption have become major challenges for organisations. To address this need, the i2CAT Research and Innovation Centre and Blueroom Innovation have launched the MAIA (Autonomous Model of Distributed Artificial Intelligence to improve Sustainability and Safety in AI environments). The initiative is part of the Corporate R&D projects programme and is supported by ACCIÓ.

MAIA’s main objective is to create an autonomous, collaborative, sustainable, and secure multi-agent AI system based on open-source language models. At i2CAT, we lead deep technological research to design the architecture of these digital agents and equip them with a fundamental layer of governance, cybersecurity and privacy. The aim is to enable AI to make complex decisions autonomously without ever compromising information confidentiality.

State-of-the-art technology for digital sovereignty and sustainability

To enable secure, efficient management, i2CAT integrates state-of-the-art cryptography tools such as Proxy Re-encryption, Zero-Knowledge Proofs (ZKP), and decentralised identities. This set of technologies allows AI agents to exchange information and execute tasks in a fully private, traceable and auditable manner.

Furthermore, the solution is designed to be implemented on-premises or in Edge environments — that is, on devices with limited resources or servers located within the same organisation. In this way, the models run directly on the company’s or institution’s infrastructure, preventing critical data from leaving the corporate environment and significantly reducing the energy impact.

Direct impact on the territory and industry

Thanks to its modular architecture, the solution developed within the framework of MAIA will adapt to the needs of strategic sectors that manage highly sensitive data:

• Industry 4.0 and logistics: optimisation of complex industrial processes and improvement of on-site energy management.
• Public Administration: support tools for public sector professionals and applications for smoother and more secure interaction with the public.
• Digital health and energy: environments where the protection of patient or electrical grid data is critical.

With this project, at i2CAT we reinforce our commitment to rigorous technological innovation with a high social and economic impact, putting advanced technology at the service of a safer, more sovereign and sustainable digital society.