Definition
Epistemic security is an interdisciplinary field — emerging in the late 2010s at the intersection of cognitive science, security studies, philosophy of information, and AI safety — concerned with systemic threats to collective and individual capacities for knowledge formation. It treats the human knowledge ecosystem as critical infrastructure on par with electrical grids, water systems, or financial networks: pervasive, indispensable, and vulnerable to deliberate attack, structural decay, or emergent failure modes.
Foundational treatments include the Alan Turing Institute's Tackling Threats to Informed Decision-Making in Democratic Societies (Seger, Avin, Pearson et al., 2020), which defines epistemic security as the conditions under which a society can reliably acquire true beliefs, update on evidence, and aggregate distributed knowledge. The framework distinguishes:
- Microepistemic threats — attacks on individual cognition (manipulation, gaslighting, persuasion exploits)
- Mesoepistemic threats — attacks on knowledge-producing institutions (academia, journalism, judicial process)
- Macroepistemic threats — attacks on the ecosystem itself (platform governance failure, AI-generated content saturation, epistemic monoculture)
The field's central insight is that truthful information is not enough: a population requires functioning epistemic infrastructure — credentialing systems, source-diversity, time-tested heuristics, calibrated trust networks, institutional checks — to convert information into knowledge. Adversaries can degrade this infrastructure without producing a single false statement, e.g., through volume-attack ("firehose of falsehood"), credentialing-attack (delegitimizing experts), or attention-capture (drowning signal in engagement-optimized noise).
Epistemic security overlaps with but is distinct from cybersecurity (which protects systems), information security (which protects data), and content moderation (which polices specific items). Its proper object is the integrity of the knowledge-formation process itself.
Context
Within the Prismatic Platform, epistemic security is the organizing frame that unifies otherwise disparate subsystems:
- The Kompendium (/kompendium/) is itself an epistemic-security artifact: a taxonomy of attack techniques against individual and collective cognition, designed for use by qualified clinicians, researchers, and security professionals.
- NABLA Infinity is a platform-level epistemic-security pipeline: source-diversity invariants, falsifiability tracking, provenance audits, and multi-modal cross-checks together implement a defense-in-depth posture against the EA technique catalog.
- The Decision Engine treats epistemic-process integrity as a first-class confidence signal: a decision is only as credible as the epistemic process producing its inputs.
- Color Teams operationalize the field's recommendations: Red Team simulates attackers, Blue Team hardens defenses, Purple Team integrates findings.
- The Info-Ops, Deepfake, Firehose of Falsehood, and Deplatforming entries each address specific subdomains of the broader epistemic-security field.
Epistemic security is the macro-frame within which the Kompendium's micro-level technique catalog operates. Defending individuals against EA techniques is necessary but not sufficient; the surrounding ecosystem must also be defended.
See Also
- Info-Ops — Military-grade information operations
- Deepfake — Synthetic-media threat surface
- Deplatforming — Platform-level defensive intervention
- Firehose of Falsehood — Volume-attack technique
- Epistemická anihilace — Microepistemic harm category
- NABLA Infinity — Platform epistemic pipeline
- Kompendium — Full Czech-language taxonomy
Connect & Contribute
Created by Tomáš Korcak (korczis) | Open Source under GHL