INESC TEC leads project to facilitate the adoption of Generative AI in industry

INESC TEC is leading a project aimed at making Generative Artificial Intelligence (AI) more accessible, efficient, and applicable within industrial contexts.

With funding of nearly €160K granted by the Foundation for Science and Technology (FCT), the project S2GAIIA – Specialized and Sustainable Generative AI for Industry Applications focuses on developing specialised, sustainable, and human-centred models, with a direct impact on the modernisation and digitalisation of the industry sector.

Although large generative AI models, e.g., GPT-4, have shown great potential, they still face limitations in complex industrial scenarios due to their size, computational costs, and lack of domain specialisation. This is where S2GAIIA comes in: to develop smaller, more efficient models adapted to specific industrial tasks, promoting their integration into organisational processes and teams.

The main goals of the project include identifying promising applications for Generative AI in industry, creating datasets and benchmarks, and developing prototypes of collaborative agents inspired by digital copilots but adapted to specific industrial contexts.

“More than a purely technical approach, the project adopts a human-centred architecture, incorporating social, contextual, and organisational factors into model development. This outlook aims to ensure the effective and ethical integration of Generative AI into industrial systems, fostering collaboration between humans and intelligent agents in complex tasks,” explained Davide Carneiro, INESC TEC researcher in enterprise systems engineering and PI of the project.

The Institute leads the S2GAIIA project, which features several companies like Navigator, EFACEC, and InfiniteFoundry – which will provide concrete use cases for prototype testing and validation.

The project will contribute to the democratisation of Generative AI, accelerating the digital transition of Portuguese industry through more accessible and sustainable solutions; at the same time, it may inspire new approaches to human-machine collaboration, based on Generative AI adapted to the real needs of organisations.

The researcher mentioned in this news piece is associated with INESC TEC