The Forensic Alternative — Est. 1862
Independent Verification for the Age of Synthetic Media
“C2PA tells you what the creator claims. AFIP tells you what the evidence shows.”
“C2PA tells you what the creator claims. AFIP tells you what the evidence shows.”
Current media provenance systems rely on creators voluntarily attaching cryptographic metadata to their work. This self-labeling approach—the foundation of standards like C2PA and Content Credentials—has three structural weaknesses that no amount of adoption can resolve.
First, metadata is routinely stripped during distribution. When a photograph is uploaded to Instagram, X, or TikTok, the platform recompresses the file and removes embedded provenance data. The content credential exists at creation and vanishes at consumption—precisely where verification matters most. AFIP research found that six of the seven largest social platforms strip all embedded provenance metadata on upload.
Second, self-labeling proves history, not truth. A content credential attests that a specific entity claims to have created content in a specific way. It does not prove the content is authentic. The distinction between “this person says they made this” and “this content has not been synthetically generated” is the difference between a self-declaration and a forensic analysis.
Third, current provenance standards embed creator identity into content metadata, creating documented privacy and doxing risks for journalists, activists, and creators in hostile environments. The World Privacy Forum has raised concerns about surveillance potential inherent in identity-linked provenance systems.
Rather than asking creators to label their own work, FIP enables independent third-party verification of any digital content—with or without metadata, with or without creator cooperation.
Content-derived hashes that survive compression, cropping, format conversion, and screenshot capture. Register a fingerprint at creation; recover it anywhere in the distribution chain.
Independent detection engines for text, image, audio, and video that identify synthetic generation through statistical, structural, and contextual analysis layers.
Cryptographically signed analysis reports in machine-readable format. Verifiable by any third party. Trust the evidence, not the institution.
Federated fingerprint registry enabling content lookup across the entire distribution chain. Find the provenance of any content, even after metadata has been stripped.
AI text detection through linguistic pattern analysis, statistical modeling, and content provenance verification. Our research covers GPT, Claude, Gemini, Llama, and emerging language models.
Spectral analysis, vocal biomarker identification, and neural pattern recognition for synthetic speech detection. Analyzing 14 vocal biomarker categories across multiple synthesis architectures.
Pixel-level examination, GAN artifact detection, and facial consistency analysis for deepfake identification. Source model attribution for synthetic imagery across Midjourney, DALL-E, Stable Diffusion, and Flux.
AFIP Verify applies multi-modal forensic analysis to determine whether content is authentic, AI-generated, or manipulated—without requiring metadata or creator cooperation.
Detect AI-generated text from GPT, Claude, Gemini, Llama, and other LLMs through linguistic forensics and statistical analysis.
Identify deepfakes, GAN-generated images, AI art, and manipulated photographs through pixel-level forensic examination.
Authenticate voices and detect synthetic speech, voice cloning, and audio manipulation through spectral and biomarker analysis.
Detect face swaps, lip-sync deepfakes, and AI-generated video through temporal coherence and frame-level forensic analysis.
Newsrooms, universities, platforms, and government agencies use AFIP certification to demonstrate forensic verification capabilities to their audiences. Certified organizations gain access to the AFIP verification API, institutional-grade analysis tools, and the right to display the AFIP Certified mark.
Think of AFIP certification as a UL listing for content integrity—an independent mark that says “this organization verifies what it publishes.” Three tiers serve different institutional needs, from editorial newsrooms to forensic laboratories to national security agencies.
arrow_forward Explore Certification Programs
A forensic examination of C2PA adoption, platform metadata stripping, and why self-labeling falls short of its verification promises.
How social media platforms handle—and destroy—embedded provenance data. Platform-by-platform analysis with perceptual hash survival rates.
Regulatory mandates for AI content labeling are growing, but enforcement requires forensic verification. Analysis of the gap between policy requirements and technical capability.
“RAND researchers identified that metadata-based provenance approaches face "fundamental challenges in adversarial environments" where content creators have incentive to mislead, recommending forensic analysis as a complementary verification layer.”
“Microsoft’s provenance research acknowledged that "current watermarking and metadata approaches cannot alone solve the provenance challenge," citing platform processing, adversarial attacks, and the open-source model problem as structural limitations.”
“The World Privacy Forum documented surveillance risks inherent in identity-linked content provenance systems, noting that "tying creator identity to every piece of content raises serious concerns for journalists, activists, and citizens in authoritarian contexts."”
Founded in 1862 as the Army Medical Museum, the Armed Forces Institute of Pathology became the world’s gold standard for forensic examination. From the JFK assassination to the Challenger disaster to 9/11 victim identification, AFIP set the standard for evidence-based truth in the most consequential investigations of the modern era.
Today, AFIP carries that forensic legacy forward into the digital age—applying the same rigorous, evidence-based methodology to the new frontier of synthetic media, AI-generated content, and digital provenance.
history Explore Our Heritage