In a bid to enhance the trustworthiness of Artificial Intelligence (AI) systems, the THEMIS 5.0 Consortium congregated in Lyon, France at the end of April, to explore progress made during the first 6 months of the project and address critical challenges unearthed through initial research. The meeting, hosted by IBM, saw the consortium, delve into the complexities of defining AI trustworthiness, how to apply a risk management approach to its assessment, including legal and ethical elements. The aim is to co-create an innovative solution to empower users in assessing trustworthiness of AI decisions for themselves, enabling them to make informed choices about using the results, as well as providing constructive feedback to AI developers for service improvement.
A notable outcome from the first half year of works were findings from the initial co-creation workshops managed by the Danish Board of Technology (DBT), the Institute of Philosophy and Technology (IPT) and Trustilio. These workshops were tailored to the project use cases - media, health and port logistics - and were held in Austria, Bulgaria, Denmark, Greece, Lithuania, Poland, Spain and the Netherlands. Data from the workshops is currently being analysed and refined into specifications for the project's technical partners in an official report will be made available in June. An overview of the process, techniques and results will be made available on the THEMIS 5.0 website at that point.
Following on from the co-creation work, the Consortium also engaged in a spot of group-work themselves, brainstorming together what will be needed for the Living Lab phase of the project when the socio-technical solution is agiley built with user feedback.
Moreover, use case partners had the opportunity to share their specific needs, contributing valuable insights to the ongoing discussions."This meeting was an important milestone in our journey towards establishing trust in AI systems. By bringing together expertise from across Europe, we are making significant strides in defining AI trustworthiness and developing practical solutions," said Hara Stefanou, the THEMIS 5.0 Coordinator, from Maggioli.
More information regarding the technical discussions and outcomes of the meeting will be made available here soon indedicated blog posts. To ensure you don't miss out on valuable content you may subscribe for updates on our website www.themis-trust.eu