JUST Data Annotation


MTF Labs has partnered with the Industry Commons Foundation on a new initiative to create JUST Data annotation standards. The project is funded by the Swedish Innovation Agency Vinnova and sets out to create standards for data practices that aim to be Judicious, Unbiased, Safe, and Transparent.

The project recognises the pervasive impact that data has on artificial intelligence (AI) and machine learning (ML) systems, where biased data can reinforce and multiply the effect of inequalities. By focusing on the annotation of data, JUST Data seeks to illuminate and address biases at their source, providing a more ethical foundation for Responsible Research and AI training.

The plan is to establish, test, iterate and ultimately standardise a metadata annotation framework for all industry and research datasets, regardless of subject or content. Data annotation is not just about tagging but involves a deep consideration of the data’s origin, context, and potential biases. This process is crucial for developing industrial and AI systems that are fair and equitable, and this is applicable across all industry sectors.


Collaborative Testbeds

MTF’s JUST Data Labs will take place at GoCo Health Innovation City in Gothenburg (June 2024) and Linköping Science Park in collaboration with Linköping University and Innovative Materials Arena (September 2024). These testbeds serve as hands-on experimental testbeds for both the JUST annotation practices and the JUST metadata interface being developed by the Industry Commons Foundation, bringing together diverse industries and experts using MTF Labs’ methodology of participatory innovation.

Building on our research for the European Open Science Cloud (Expanding EOSC, 2020), the JUST Data initiative enhances and supports the FAIR data principles (that data should be Findable, Accessible, Interoperable, and Reusable). The intention is to add layers of ethical considerations that are not about the data themselves, but about the practices of the researcher, data manager and data owner.

While FAIR principles ensure data is managed in a way that enables it to be shared and used efficiently, JUST principles ensure that this use is ethically responsible, acknowledging and mitigating biases, including gender biases, to promote fairness and inclusivity in AI development and application in all industrial domains.

Through the JUST Data initiative, our ambition is to create standards that guide the ethical annotation of data and foster a broader understanding and adoption of these practices within industry. The JUST  project underlines the critical role of responsible data management in building AI systems that serve all segments of society equitably. By addressing biases at the data level, we can pave the way for more just and inclusive technological futures.

Photo credits: ThisisEngineering RAEng, National Cancer Institute and Andrea Magaš-Pavušek

7A Centralen
Vasagatan 7
111 20 Stockholm


Privacy Policy