Objective
Artificial intelligence (AI) and data services have experienced rapid and remarkable advancements in recent years, driving breakthroughs across services, capacities and technologies across various domains. While automation has made significant progress, it has also introduced risks associated with untested, faulty or unsecured data and AI solutions. With this in mind, the EU-funded DataPACT project aims to develop tools, methodologies and approaches to help data and AI pipeline developers create ethical, compliant and environmentally sustainable solutions. The project aims to deliver three key resources: a compliance toolbox, a compliance framework and a data/AI pipeline toolbox. Together, these will form a comprehensive development ecosystem, providing user-friendly tools to ensure quality and sustainability in data and AI operations in various use cases.
SINTEF's role in the project
SINTEF is the coordinator of the project and leads the development of a compliance-aware data pipeline toolbox to support the entire operational lifecycle of data/AI pipelines, covering design, simulation, deployment, execution, monitoring, and traceability. Specifically, this toolbox aims to (1) integrate compliance as a first-class citizen in all the design and runtime aspects of data/AI pipelines; (2) provide an environmentally friendly run-time system to data/AI pipelines through energy monitoring, carbon footprint estimation with optimised efficiency and quality trade-offs; and (3) successful application in fine-tuning foundation models, data sharing in data spaces and retrieval augmented generation.
Project website
Funding
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No 101189771.