The Ethics of Amplification

Social media platforms’ AI systems learn from user data to predict what content will keep users engaged. This content is subsequently amplified, increasing its visibility. Over the past years, many critics have claimed that such an approach to algorithmic amplification is morally impermissible. For some, the wrong inheres in its manipulative character. For others, the wrong is that of making the users themselves worse off (e.g., by squandering their time). And for others, the wrong lies in perpetuating the injustice of surveillance capitalism. I will argue that these existing diagnoses fail to capture what is genuinely objectionable – and what specifically needs reform – in contemporary algorithmic amplification systems on social media. I argue that platforms have a civic duty to support the successful development and exercise of citizens’ moral and civic capacities; this duty requires diversification of the criteria governing the amplification of politically salient content, and it is this duty that platforms currently breach.