Election Interference by Algorithm: When Platform Choices Shape Electoral Outcomes

Algorithmic manipulation of political attitudes during campaign periods potentially constitutes a form of election interference operating below traditional oversight mechanisms designed for foreign influence and illegal coordination rather than corporate algorithm optimization.
The research specifically examined over 1,000 users during the 2024 US presidential election, finding that one week of altered feeds produced substantial polarization shifts. These effects occurred during the precise period when citizens were forming opinions and making decisions that would determine electoral outcomes.
If algorithms can significantly shift political attitudes within campaign timeframes, they potentially influence election results beyond whatever effects the content itself might have. A platform amplifying content that increases animosity toward one party’s supporters might indirectly benefit the opposing party by making fence-sitting voters less willing to identify with the targeted group.
Traditional election interference involves foreign governments, illegal coordination, or campaign finance violations. But algorithmic interference operates through corporate decisions about engagement optimization that face minimal oversight. Companies can amplify content affecting electoral outcomes while claiming they’re merely maximizing user engagement rather than deliberately influencing politics.
Addressing algorithmic election interference might require new regulatory frameworks. Campaign periods might require special algorithmic constraints or transparency. Platforms might be prohibited from making algorithmic changes during defined pre-election periods. Or algorithms might be required to operate identically for all politically-relevant content regardless of engagement patterns. Whether democracies will develop such frameworks before algorithmic influence fundamentally compromises electoral integrity remains an urgent open question.