At the 4th conference of the TAILOR Network of Excellence Centres on Trustworthy AI entitled “Trustworthy AI from Lab to market” in Lisbon, Mattheos Fikardos presented joint research with Katerina Lepenioti, Dimitris Apostolou and Gregoris Mentzas on using an LLM-based multi-agent framework for supporting trustworthy AI covering the whole AI lifecycle in order to facilitate conformity with the EU AI Act. This work exploits  open-source methods and toolkits for accuracy, robustness, bias mitigation, etc.

Research within ICCS-NTUA funded by the EC funded projects Themis 5.0 and Faith.