Developing a learning framework for Trustworthy AI: the beginning
Within the Trustworthy AI project, our team at UMU is tasked with leading the first Intellectual Output: developing a Teaching Framework for Artificial Intelligence (AI) in Higher Education. The framework will be based on the framework created by the European Commission and other highly relevant stakeholders (https://ec.europa.eu/futurium/en/ai-alliance-consultation) on what reliable and ethical AI should look like. This framework will present an integrated approach to teaching AI, including its social, ethical, and economic aspects, but also the principles and learning strategies needed for higher-education students to obtain all the necessary competencies for the development of trustworthy AI.
An important goal of this framework is to give a strong basis for the development of high-quality educational material in the remainder of the project. For this reason, we are taking a highly focused approach around three specific research questions:
What competencies need to be acquired by students in higher education so that our future workforce is equipped to implement the European Guidelines for Trustworthy AI?
How to teach and evaluate these competencies in the context of higher education?
What is the current focus of AI education with respect to trustworthy AI? Which resources are available, and which aspects are missing?
In our approach to answer these questions, we aim to take a broad perspective with a multi-disciplinary outlook. We believe an inclusive perspective is fundamental, as the Trustworthy AI principles endorsed by the EU include many inter-disciplinary concepts that require insights into how to develop social and ethical reasoning skills tackled in fields such as medicine or the social sciences. Furthermore, the EU includes a variety of perspectives and educational systems brought by its multiculturality: it is, therefore, important to capture this richness in the developed educational framework.
With these factors in mind, our UMU team is undertaking two main tasks that will inform the Teaching Framework. The first, a systematic literature review, aims to reveal which competencies are needed to equip higher education students with skills enabling the development of trustworthy AI. This review has the additional goal of studying which educational approaches are effective for the teaching and assessment of such competences. The second task is dedicated to obtaining a multi-cultural and multi-disciplinary perspective into the current state and needs of education concerning trustworthy AI. For this purpose, we developed an interview protocol to gather the insights of 15 to 20 educators, policymakers and other stakeholders around the EU. With the help of our wonderful partners, who will each conduct a significant portion of the interviews, in the next few months we will gather the invaluable insights of stakeholders of a variety of backgrounds and locations. This information will then be aggregated and analysed, resulting in a comprehensive outlook into the status of trustworthy AI education in the EU.
As this work progresses, we are looking forward to creating a comprehensive Learning Framework grounded in established knowledge and answering current needs. We will keep you updated!