Linguistic Signal Integrity Node Active

Epistemic Security
Infrastructure.
Making Rhetorical Manipulation Visible.

Project Integrity analyzes public speech to identify rhetorical manipulation patterns — before they normalize hostility toward entire groups.

Taxonomy Model V0.4 Stable
Case Bases 13 Samples
Classifications 5 Core Modes
Explore Taxonomy No Editorial Agenda

The Problem

Political manipulation has always existed. What's new is the scale. Coordinated rhetoric, repeated patterns, escalating language — these techniques exploit the same cognitive shortcuts in all of us, regardless of education or ideology. No society has reliable infrastructure to detect this systematically, in real time, before it normalizes hostility toward entire groups.

Our Approach

Project Integrity builds AI tools that analyze public speech for rhetorical manipulation patterns — not to judge politicians, but to make the techniques visible. Transparent methodology, open data, no editorial agenda. The same way financial audits make the numbers public — we make the rhetoric public.

Rhetorical Taxonomy V0.4

Detection Context

Same Words. Different Context.
Different Classification.

5 of 13 · always includes min. 1 NOT DETECTED + min. 1 LOW/RHETORICAL

Pilot Dataset — IL-V0.4

AIgator.live — Knesset Rhetorical Corpus

The first algorithmic audit of Israeli parliamentary speech. Built on the HaifaCLGroup/KnessetCorpus (HuggingFace, CC BY-SA 4.0), cross-referenced with live Knesset OData API.

Periods covered

2021–2022 elections
· Judicial reform crisis 2023
· 2026 elections

Actors tracked
10
Detections
4,599
Methodology Open source, CC BY-SA 4.0
Explore Pilot Dataset → aigator.live