Bot and Sockpuppet Detection

Definition

Bot and Sockpuppet Detection examines distinguishing deceptive automated or multiply controlled identities from legitimate pseudonymity and automation within the broader domain of covert, deceptive or exploitative attempts to shape beliefs, behaviour, attention or collective action.

Why this matters

Bot and Sockpuppet Detection can materially affect autonomy, safety, public trust, market integrity, community cohesion and institutional decisions. Poorly designed interventions can suppress legitimate speech, entrench bias or create false confidence.

Research questions

Under which conditions can distinguishing deceptive automated or multiply controlled identities from legitimate pseudonymity and automation be measured or improved reliably, and how do effects vary by platform, population, context and intervention?

Hypotheses

A transparent, context-aware approach combining provenance, behavioural evidence and accountable human review will improve detection precision, recall and user harm more than single-score, content-only or opaque automated approaches.

Proposed methods

coordinated-behaviour analysis; graph analytics; campaign reconstruction; synthetic-media forensics; controlled persuasion experiments; red-team exercises; threat intelligence; expert review; affected-user interviews; reproducibility testing; methods adapted specifically to Bot and Sockpuppet Detection

Stakeholders and beneficiaries

election authorities; civil society; platforms; journalists; security researchers; public agencies; brands; communities; targeted populations