BIAS Consortium and Advisory Board meeting in Porto

On the first day, accomplishments in each of the work packages were presented. The excitement peaked on the second day when discussions turned to the next steps: “What should we truly do to mitigate algorithmic bias in the labor market? What can be considered bias and what should not? How should bias be tackled on a legal level?” And importantly, “How should the core outcomes of the project be communicated to the world?” Fortunately, unlike a decade ago, stakeholders (e.g., citizens, researchers, and companies) are aware of algorithmic bias and its potential impact on societies and workplaces. Still, the question of detecting and addressing such bias remains open to many. During the meeting, both the partners and the advisory board agreed that a common list of definitions should be established to enable a unified discussion of bias in recruitment from technical, legal, and social perspectives. All stakeholders should be empowered by the debiaser toolkit and methodologies to check their systems for any algorithmic bias.

 

For the development and deployment of such a novel toolkit, Digiotouch set up a multi-cloud federated infrastructure. It hosts the anonymised data from Farplas in a GDPR-compliant manner enabling testing and validation of the developed tools. During the meeting, discussions took place to augment the resources of the federated infrastructure to support the increasing computational needs of AI researchers.

 

Another hot topic of discussion was the upcoming Summer School on Bias in AI, being organized by Digiotouch. It will take place in Tallinn in April 2025. Two Keynote speeches from the BIAS partners BFH and NTNU and an industrial Keynote from Google were confirmed. Digiotouch as well as the project partners warmly invite researchers, and techies including software developers and data scientists, HR professionals, and all other stakeholders to the summer school. A trustworthy AI helix event is also being collocated with the summer school.

 

On a concluding note, BIAS partners have created national labs comprising communities of practitioners, employees, HR managers, researchers, and AI specialists with a special focus on underrepresented communities, not only to seek feedback through co-creation but also to empower them with the best practices on algorithmic bias detection and mitigation. Join the Estonian National Lab to stay updated with the project activities and best practices.