GenAI fraud makes zero-knowledge proofs non-negotiable By Jarek Sygitowicz, Co-Founder and Chief Strategy Officer of Authologic By now it’s something of a cliché to say the rapid rise of generative AI has fundamentally changed the threat landscape of the internet. Nevertheless, it has, and the point bears repeating. Synthetic identities, deepfakes and AI-generated documents are no longer edge cases but the go-to tools of fraud. The more convincing generative models become, the faster they sideline traditional methods of digital verification like uploading photos of IDs, sharing personal data or performing visual checks. The more data people share, the more material fraudsters have to exploit. Identity verification methods still depend on oversharing. Few flaws are more pressing. Zero-knowledge proofs (ZKPs) offer a way out of this trap. Instead of asking users to reveal sensitive information to prove eligibility, ZKPs allow them to prove a fact about themselves without disclosing the underlying data. When AI can fabricate a face or a document, a photo becomes largely useless as a verification method. More than that, ZKPs limit data exposure. Once viewed as a privacy improvement, data minimisation is now a security imperative. Oversharing as a systemic vulnerability Most online verification flows today unintentionally collect hordes of unnecessary information. Age checks often require a full photo of a government-issued ID, collecting addresses, height and weight, and eye color in the process. Address verification exposes home locations. Identity checks routinely capture names, document numbers, and biometric images, all stored beyond the reach and control of the end user. When fraud required human effort and manual forgery, this model provided adequate protection. GenAI has made this model dangerously outdated almost overnight. A single leaked ID image can be reused and altered at scale. AI systems trained on real documents can generate convincing counterfeits that outpace any