Edition 2
The Invisible Programme Controller
How recommendation systems organise attention, shape exposure and influence children, politics and public life.
From the Editor
“The interesting questions are often around the model.”
The debate about social media often begins with content and ends with bans. This edition asks a different question: who decides what appears next? Recommendation systems now perform a role once held by programme controllers, but they make millions of individual choices continuously and with limited public accountability. The articles that follow examine what these systems demonstrably change, what they may influence and where the evidence does not support a causal claim.
Dr Gary Marrs PhD Editor, Beyond the Model
Recommender systems do more than help people find material. They order attention. They decide which item is prominent, which voice is repeated and which possibility remains unseen.
This edition begins with children because the conflict is especially clear: online services can create real harm and provide real developmental value. It then follows the same mechanism into politics and violent extremism, carefully separating changes in exposure from changes in belief and action.
The closing framework asks how services can govern recommendation in the public interest without turning every feed into a prescribed curriculum.
Inside Edition 2
In this edition
Eight connected questions, designed to be read in sequence or entered at the point most relevant to you.
- 01What Do We Mean by Social Media?A blanket label conceals important differences between public feeds, video libraries, messaging services and school communities.
- 02What Would an Under-16 Ban Actually Require?Age restrictions sound simple until identity, privacy, proportionality and circumvention become part of the design.
- 03When Online Video Is Developmental SupportSome children use repeated language, songs and situations as material for communication and learning.
- 04The Algorithm Has Become the Programme ControllerRecommenders perform an editorial function at immense scale, but their choices are governed differently from a broadcast schedule.
- 05Engagement Should Be the Tool, Not the GoalAttention can serve learning, creativity, connection and entertainment without becoming the final measure of value.
- 06Does Your Algorithm Make You More Like Yesterday?Feedback loops can make inferred preferences increasingly influential without proving that the person underneath has changed.
- 07Can a Recommendation Path Become a Radicalisation Path?Online systems can assist discovery, repetition and community formation, but the leap from exposure to extremist violence is not a settled causal chain.
- 08A Public Interest Framework for Recommender SystemsEight questions for governing systems that organise attention across childhood, politics and public life.