Opening the black boxes of algorithms, a serious game

At a time when AI is creeping everywhere in our daily lives, the public administrations have not escaped the AI-mania. This has resulted in complex algorithms being used in various fields that were previously managed by humans, such as education. A notable example from the UK is the algorithm used to grade A-Level students in 2020, when the cancellation of exams due to the pandemic meant that alternative methods had to be found. This case clearly demonstrates the use of complex algorithms to replace teachers and reshape access to higher education. Although the committee responsible for assessing the algorithm approved its development, public controversies and protests ultimately led to its cancellation. Was the pupils’ selection process skewed by the algorithm? Could it be accused of being unfair and augment social inequalities? How to address technological tools that remain black boxes to decision-makers? Rather than turning it into a traditional academic paper, I used the lessons learned from this case study to develop a serious game aimed at identifying the impact of this algorithm on social inequalities. My goal is to playfully explore the challenges of understanding how algorithms work and their impact on society and policy-making. I suggest organising a one-hour session with TA practitioners to discuss the accessibility of such technology, considering not only secrecy, but also understandability when assessing complicated systems. The discussion of the algorithm used to grade English pupils in 2020 is particularly pertinent to this EPTA meeting, as it encompasses both the current trend towards everything AI – questioning its impact on pupils’ selection – and the preferred methods to open such “black boxes” and foster debate about societal acceptance and policy relevance.