Research map / Large language models / AI and society
Merging AI Incidents Research with Political Misinformation Research: Introducing the Political Deepfakes Incidents Database
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Introduces the Political Deepfakes Incidents Database (PDID), a researcher-coded, publicly hosted collection of U.S.-focused political deepfakes and cheapfakes (videos and images) linking AI incidents research with political misinformation research.
Problem
Existing deepfake databases focus on technical detection and do not comprehensively capture politically salient deep- and cheapfakes in the wild, making systematic study of their prevalence, trends, and impact difficult.
Approach
A codebook-driven incidents database that combines AI incident taxonomies (AIID, CSET, AVID) with political science and communication constructs (communication goals, harm types, policy sectors, framing, and Narrative Policy Framework narratives) to code real-world political deepfakes with metadata, verification, and real-world connections.
Key results
- Data collection began June 2023 and is ongoing, with incidents dating back to 2017, hosted on AirTable.
- Database covers videos and images (audio excluded initially) primarily from the U.S. and is described as a publicly available resource.
Limitations
- Limited to English-language sources and U.S. focus, with restricted regional and linguistic coverage
- Manual coding of variables; harm and communication-goal coding involves subjectivity and inference
- Excludes audio-only content and highly incendiary sexual or violent content
Experimental setup
Datasets: social media posts popular news websites
Builds on: AI Incidents Database CSET AI harms taxonomy Narrative Policy Framework
Closest papers by meaning
- Quantifying Political Polarization through the Lens of Machine Translation and Vicarious Offense (AAAI 2024 · similarity 0.82)
- Lessons for Editors of AI Incidents from the AI Incident Database (AAAI 2025 · similarity 0.81)
- The Power of LLM-Generated Synthetic Data for Stance Detection in Online Political Discussions (ICLR 2025 · similarity 0.81)
- Political Bias Prediction Models Focus on Source Cues, Not Semantics (AAAI 2025 · similarity 0.81)
- TriDF: Evaluating Perception, Detection, and Hallucination for Interpretable DeepFake Detection (CVPR 2026 · similarity 0.81)
- Beyond Detection: Exploring Evidence-based Multi-Agent Debate for Misinformation Intervention and Persuasion (AAAI 2026 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.