Identity In The Era Of Deepfakes, AI And Mass Biometrics Exposure

The talk opens with a Ponzi scheme advert that used a deepfake of a well-known public figure, then asks what changed to make identity abuse so common. The speaker links several trends: social feeds that reward snap judgments, ransomware groups that added data leaks to encryption, cheap deepfake and voice-cloning services sold underground, and public playbooks for manipulating opinion.

The core argument is that biometric traits such as faces, voices, fingerprints, irises and palm shapes leak through everyday posts, often without anyone noticing. Examples included high-resolution politician photos on a government website and a fingerprint visible on a wine glass posted online. In small hardware experiments, an iris sensor accepted a crudely drawn eye, and some images found online were enrolled successfully. The speaker argued that biometrics only work as a trust anchor when the verifier controls the sensor. Underground markets were described as selling real identities, deepfake tools, and guides for bypassing remote account checks, alongside virtual kidnapping scams, blackmail, and attacks on voice-controlled devices.

The talk closes by treating elections, summits and concerts as short windows in which fabricated media can cause outsized damage. Its advice is to limit where biometric images and recordings are shared, and to avoid using fingerprints for financial approvals. For practitioners, the takeaway is that biometrics should be paired with a second factor rather than trusted alone.