Autonomous robotics, agentic Artificial Intelligence (AI) for engineering design and simulation, clinical datasets, and mobility are the four areas in which researchers, start-ups, small and medium-sized enterprises (SMEs), and academic teams from across the European Union and associated countries are invited to showcase the potential of their generative AI innovations. The competition, organised as part of the AI-BOOST consortium, forms part of a wider portfolio of highly scalable open innovation initiatives in AI. Designed to attract talented researchers and innovators from across the EU and associated countries, the challenges aim to advance scientific progress in key AI sectors.
More specifically, the AI Challenge Competition seeks to help projects progress from Technology Readiness Levels (TRLs) 2-3 to TRLs 4-5; participants will compete for cash prizes ranging from €28,5K for teams reaching the Spark phase to €100K for the winners of each of the four main challenges in the Advance phase. Before reaching those stages, participating teams will be assigned to one of four challenge areas and supported by the industrial partner leading that challenge. Further details are available on the competition’s official website.
The first challenge, led by the Consorzio Intellimech and JOiNT LAB, focuses on Generative AI-Based Natural Language Mission Creation for Autonomous Agricultural Robots. Participants will be invited to democratise robot programming by using advanced Vision-Language-Action (VLA) models. The goal is to develop a system capable of translating natural language into executable robot commands, initially to improve productivity in vineyards but with the flexibility to support a broader range of agricultural field operations.
The second challenge, led by SIAD Group, focuses on Agentic AI for Automated CAD Generation and Autonomous Simulation. Participants are tasked with simplifying the design and simulation of industrial piping systems, particularly compressor pipe routing in skid-mounted systems. Proposed solutions should be capable of translating natural language requirements and CAD data into editable parametric models. This would automate complex engineering tasks, like extracting geometric constraints and designing piping networks, significantly reducing manual engineering effort.
The third challenge, led by the European Federation for Cancer Images (EUCAIM), explores the use of Generative AI to Improve the Quality and Representativeness of Clinical Imaging Datasets. Such datasets are often incomplete or unbalanced. Participants are expected to develop solutions capable of generating synthetic patient cohorts – groups of individuals who share common characteristics, such as age or exposure to a particular event, and are monitored over time to study how they evolve – to address data gaps and harmonise information across different imaging conditions. The proposed systems should identify key demographic and clinical characteristics to ensure that the synthetic data remain realistic and representative of real-world populations.
The fourth and final AI Challenge Competition, led by Siemens Industry Software NV, centres on Generative AI for the Automated Generation of Test Cases from Accident Databases and Safety Standards. The objective is to strengthen the safety assessment and validation of autonomous driving systems by using generative AI to automate the creation of simulation scenarios. Drawing on accident reports, visual data and international safety standards in structured formats, the proposed solutions should replace slow, traditional workflows with a scalable and consistent process. By linking real-world data with regulatory frameworks, this challenge could provide engineers and safety experts with comprehensive new sets of test scenarios, improving both the efficiency and accessibility of autonomous vehicle safety assessments.
As part of the competition, five winners will be selected in each challenge area during the Spark phase. They will then take part in an intensive five-month acceleration and validation programme. At the end of the programme, the finalists will present their solutions publicly at a showcase event in Brussels in February 2027.
News, current topics, curiosities and so much more about INESC TEC and its community!
