Deepfakes and voice cloning used in cyberattacks to evade identity verification
Cybercriminals reportedly use AI-generated deepfakes and voice cloning to bypass biometric identity verification in financial institutions, enabling account takeovers and fraud. Security researchers identified tools and darknet markets facilitating these attacks, with costs ranging from $30 for video datasets to thousands of dollars per minute for high-quality synthetic content.
What happened
Cybercriminals use deepfakes and voice cloning generated by AI to bypass biometric identity verification methods in banks and other financial institutions. Sensity identified 47 tools designed specifically to circumvent KYC (Know Your Customer) identity verification procedures.
404 Media revealed a clandestine website selling photos and videos of people to evade KYC systems, with digital datasets packaged at approximately $30 per collection and featuring people of various ages, genders and ethnicities. The darknet market also recruits volunteers in disadvantaged countries, offering between $5 and $20 per person to provide such material.
Reported impact
Cybercriminals can open accounts remotely in banks, cryptocurrency exchanges and other financial services using deepfake digital clones. This enables them to steal personal information, banking data and money from financial institutions and their customers.
Celebrities and public figures whose voices or videos are available online are especially vulnerable to impersonation through deepfakes and voice cloning.
Index analysis
The darknet market for deepfakes operates in a clandestine environment where the cost of synthetic content ranges from $300 to $20,000 per minute depending on quality and complexity. Sites in this market offer reviews and highlight 'new clones' to encourage purchases, and some offer customized datasets created on demand.
Advances in generative AI and machine learning have made it possible to create increasingly realistic synthetic multimedia, with the ability to train on unstructured data such as facial features and human movements to generate convincing false images and audio.
Governance controls
Security experts from Kaspersky emphasize the need to establish ethical guidelines for AI use and highlight that responsible implementation of AI is crucial to prevent material, psychological and reputational damage.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.