In Brief

The Challenge

ARTAI addresses the challenge of ensuring trustworthy AI recommender systems. AI algorithms embedded in applications such as Instagram, YouTube or TikTok are profoundly influential in society. While they may be beneficial in suggesting new content, there is now mounting evidence that they can present serious risk, particularly to young people. This is largely due to a lack of transparency, a drive for user engagement and high levels of personalisation. ARTAI addresses this challenge of ensuring transparency in order to enable algorithmic oversight, auditing and the protection of users online. 

The Solution

ARTAI will develop a large-scale simulation platform to provide transparency into the kind of content that is recommended to people online. The system will enable large-scale simulations of different user interests, online behaviour and interactions with recommender algorithms over time. An AI-based approach to classification will then evaluate recommendations and visualise societal trends. The ARTAI platform is designed to advance an understanding of how content is disseminated to people online, support ethics auditing and improve oversight and transparency of recommender algorithms. 

The Team 

  • Team Lead: Dr Susan Leavy, University College Dublin  
  • Team Co-Lead: Dr Ruihai Dong, University College Dublin  

Societal Impact Champion  

  • Jean Carberry, Department of Enterprise, Trade and Employme