Research
Research & Analysis
Forensic-grade analysis of AI-generated content, media provenance, and synthetic media detection across text, audio, video, and image domains.
AFIP’s provenance research examines the technical and policy foundations of content authenticity—from platform metadata handling to regulatory frameworks to the structural limitations of self-labeling approaches.
A forensic examination of C2PA adoption, the metadata stripping problem, and why self-labeling falls short of its verification promises. Cites RAND, Microsoft, Fortune, and the World Privacy Forum.
Platform-by-platform analysis of how social media handles embedded provenance data. Perceptual hash survival rates across Instagram, X, TikTok, Facebook, YouTube, WhatsApp, and Telegram.
Regulatory mandates for AI content labeling are growing worldwide, but enforcement requires forensic verification. Analysis of the gap between policy ambition and technical enforcement capability.
AI text detection methodology, linguistic forensics, statistical pattern analysis, and content provenance verification.
Spectral analysis, vocal biomarker identification, and neural pattern recognition for detecting synthetic speech.
Pixel-level examination, GAN artifact detection, and facial consistency analysis for manipulated media.
AFIP's historic expertise in tracking biological contagions, now applied to understanding information contagion.
From military remains identification to digital content fingerprinting and provenance authentication.
AFIP's pioneering veterinary pathology legacy, including the renowned Ferret Pathology collection.
A curated repository of analyzed synthetic content samples, detection benchmarks, and forensic case studies.