A team led by the GenAI Lab’s Aaron Snoswell, with William He and Jean Burgess and in collaboration with Tariq Choucair (QUT) and Damiano Spina (RMIT) has been successful in making it through to the finals of the Prosocial Ranking Challenge.
The Prosocial Ranking Challenge is a contest hosted by the Center for Human-Compatible AI at UC Berkeley, offering $60,000 in prizes to develop social media algorithms that are healthier for individuals and society. From 21 initial entries, nine finalists were chosen to proceed to the next round.
The GenAI Lab team’s approach is based on the concept of Search Result Diversification – we aim to mitigate political and other forms of polarization by exposing users to diverse opinions and content. To do this, we use a Large Language Model (LLM) to simulate personas with diverse political perspectives, rank the users’ social media feed according to each of these personas, and then combine these partisan rankings with an award-winning fairness-preserving algorithm that balances the range of opinions present in the news feed.
Finalists now have one month to make their ranking algorithm prototype into a production ready system with stringent performance, security, and privacy requirements. Five winners will then be selected to take part in a large-scale trial to evaluate the real-world performance of the algorithms.