University of Helsinki researchers develop Somekone to teach algorithm mechanics

Researchers at the University of Helsinki, as part of the Generation AI program, have developed Somekone, a testing platform designed to teach students how social media algorithms function. The tool aims to address a gap in media literacy by moving beyond basic cybersecurity and misinformation to explain the underlying mechanisms of data collection.

Somekone provides a closed environment that mimics platforms like Instagram, allowing users to browse a feed, like, comment, and react with emojis. The system tracks granular data, including how long a user pauses on an image, the length of comments, and various interactions. This data is then used to build a visual profile of the user, demonstrating how their digital footprint is constructed and how algorithms determine content recommendations.

By using a second device, students can observe the profile being built in real-time. The platform explains why specific content is suggested, such as similarities to previous interests or patterns among similar users, helping learners understand how social media bubbles and behavioral guidance are formed.