A look at OnlyFake, an underground site that claims to use “neural networks” to make highly convincing fake IDs for just $15; one ID passed verification on OKX
This piece was produced with support from The Capitol Forum. — An underground website called OnlyFake is claiming to use …
Context & Ripple Effects
This report extends an earlier arc in which AI-generated faces were used to support fake accounts posing as Americans and synthetic people were marketed for bot-driven activity online. The relevant shift here is from fabricated social identities to documents intended to clear a platform’s identity gate.
The later warning from Socure that criminals are combining social-media selfies with generative AI for more realistic IDs suggests this is not solely a document-forgery problem; it is an escalating contest between identity-proofing systems and synthetic inputs.
First-order effects
- The reported successful OKX verification puts its identity-checking process under immediate scrutiny, because a low-cost purportedly synthetic document appears to have cleared a control meant to block fraudulent access.
- OnlyFake’s claimed offering lowers the practical barrier to attempting identity fraud by packaging convincing-looking IDs as a cheap, repeatable service.
Second-order effects
- Exchanges and other services that rely on document-and-selfie verification will face pressure to tighten checks for presentation attacks, including stronger liveness and document-authenticity testing.
- Identity-verification vendors gain a clearer demand case for defenses against generative-AI-enabled fraud, consistent with Socure’s warning about AI-enhanced fake IDs.
Third-order effects
- If synthetic IDs continue to improve faster than standard onboarding checks, digital identity verification is likely to move from one-time document review toward layered, continuously refreshed risk assessment.
- For crypto platforms in particular, recurring verification gaps could deepen the legitimacy challenge exposed by the OnlyFake report, making compliance controls a more consequential competitive and regulatory differentiator.
The trend: Generative AI is turning identity fraud into a scalable synthetic-supply problem, forcing trust systems to verify not just a document’s appearance but the provenance of the identity behind it.