Algorithmic mediation is the process by which software filters, ranks, recommends, or summarizes information before it reaches a user. It appears in search results, social feeds, news aggregators, and AI assistants. In cyber security, it matters because the mediation layer can shape trust: users may see only selected sources, truncated context, or a synthesized answer that hides where the information came from.
Attackers try to exploit this layer by poisoning search results, flooding platforms with coordinated content, manipulating recommendation signals, or using prompt injection to steer an AI summary. Defenders reduce the risk with source attribution, transparency about ranking and summarization, logging and audits, abuse detection, and diversity checks that keep competing viewpoints visible. The goal is not just correct content, but verifiable content routing.