Recommender System Regulation

Recommender systems are algorithms that determine the relative order or prominence of information presented to users. The DSA requires transparency and user choice regarding these systems.

Transparency Requirements

Online platforms must clearly present the main parameters used in recommender systems, including:

• Most important criteria determining recommended content
• Reasons for the relative importance of those criteria
• Whether and how user behavior influences recommendations
• Options available to users to modify or influence recommendations

User Choice (VLOPs)

Very large online platforms must provide at least one option for users to access their services that is not based on profiling (i.e., not based on tracking user behavior across services).

This non-profiling option must be:

• Directly accessible from the interface where recommendations are presented
• Equally easy to access as profiling-based options
• Presented in clear, plain language

Purpose

These requirements aim to:

• Increase user awareness of algorithmic curation
• Provide users with meaningful choices
• Reduce filter bubbles and echo chambers
• Enable informed decision-making about information consumption