ALTAI - Assessment List for Trustworthy AI - A Web-based prototype

Dear members of the European AI Alliance,


It is with great pride that we are sharing with you one of our final deliverables of the High Level Expert Group on Artificial Intelligence on behalf of all the expert members. The work that we have undertaken during the past two years of our mandate has been extraordinary. Our group was one of the first to set the scene by presenting Ethics Guidelines for Trustworthy AI in mid-2019, putting forward our vision for what shape artificial intelligence in Europe should take. It has been noted by national and international stakeholders and encouraged us to push onwards with this vision.

Since then we have undertaken groundbreaking efforts to encourage the implementability and accessibility of our key recommendations. Through the piloting process of the Assessment List, which accompanied our key recommendations, the European Commission has ensured that the broadest possible representatives of the community are enabled to provide feedback on the revision thereof. We thank all those who have taken the time to give us their valuable insights.

Throughout the past year then, we have collected a vast array of insights, from academia, industry, start-ups, NGOs and many more, directly shaping our new, and final output: the revised and final Assessment List for Trustworthy AI, in short, ALTAI. ALTAI is the one of the first of its kind tools that bridges the gap between principles and action - outlining a vast number of crucial questions to ask throughout the development process of a Trustworthy AI system.

We are very proud of this important document which we have presented for you in a PDF format but also, in order to demonstrate the capability of the ALTAI, in a web-based tool (currently a prototype), raising awareness about trustworthy AI and the requirements necessary for it. This tool was implemented by a team of researchers at the Insight Centre for Data Analytics and School of Computer Science & Information Technology at University College Cork (Ireland), the home institution of the Vice Chair of the AI HLEG.

It is available to all and we expect that it will inform decisions made throughout the entire lifecycle of an AI system, ensuring that Trustworthy AI is developed and deployed in Europe and that users can benefit from the vast potential and promise that AI yields for all sectors.

While our mandate ends, all of us in the High Level Expert Group on Artificial Intelligence hope that the goalposts we have set for AI in Europe will remain. We very much look forward to the wider community being able to use the reports and tools we have developed, making them their own and continuing the important task of encouraging human-centric AI in the European Union.

Kind regards,

Pekka Ala-Pietilä, Chair
Barry O’Sullivan, Vice Chair