Project goal and tasks:
HumAMI aims to strengthen the European AI ecosystem by establishing a robust Physical AI framework. Physical AI refers to AI solutions that operate within and interact with real-world physical systems such as machines, robots, and human-centric cyber-physical systems (CPSs). A Physical AI embodiment is such a machine, robot, or CPS with advanced machine intelligence (AMI). A Physical AI embodiment that interacts with humans in the loop inherently requires human-inspired characteristics – adaptability, explainability, trustworthiness, ethics, and sustainability – to be present in its entire design trajectory and lifecycle to ensure safe, secure, and robust operation. To achieve this, the HumAMI framework brings together (i) edge AI, and (ii) agentic AI enabling AMI such as reasoning, learning, action, and decision-making. Transitioning towards Physical AI also requires supporting (i) cloud resources, (ii) connectivity, and (iii) applications & services. The HumAMI framework encompasses these by introducing a 6-layer conceptual framework composed of – Physical infrastructure, Connectivity, Embedded AI, Agentic AI, Cloud platform, and Applications & Services, to realize Physical AI embodiment in industrial machines, robots and human-centric CPSs.
HumAMI framework is realised by (i) developing fundamental innovative disruptive technologies (IDTs) (open science objective openAMI), (ii) developing integrated technology solutions (ITSs) bringing multiple IDTs together (cooperation objective coopAMI), and (iii) demonstrating and validating IDTs and ITSs in real-world technology demonstrators (RTDs), fostering industrial adoption (adoptAMI objective) in mobility, energy, digital industry, health, agrifood, and digital society. The HumAMI initiative will significantly enhance European industrial competitiveness, promote the development of ethical AI, and secure Europe’s leadership in the global Physical AI landscape, balancing technological advancement with core societal values.
Tasks to be performed by EDI:
EDI is responsible for the development of IDT9: “CROSS-multi-modal interactive perception, fusion and learning”; IDT10: “Gen-AI based natural language instruction parsing and high-level planning for agent actions”; and IDT11: “Vocabulary of skills for robotic control or action”, which will be integrated, under the leadership of PAL Robotics, into RTD10 (“Socially pertinent robots for interaction and collaboration with humans”). EDI also leads Work Package 4 (WP4): “Agentic and Physical AI SW enablers”.
HumAMI is supported by the Chips Joint Undertaking and its members, including the top-up funding by National Authorities under Grant Agreement No 101252558.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them.
Participating scientists
