AI vishing is voice phishing that uses synthetic or AI-generated speech to imitate a trusted person, such as a bank employee, help desk agent, executive, or family member. Attackers use it to sound convincing, create urgency, and persuade the target to reveal passwords, one-time codes, or payment details, or to approve a fraudulent action.
This matters because voice is often treated as more trustworthy than text, especially in support and recovery workflows. In real attacks, AI vishing can be combined with phishing emails, fake login pages, and mobile abuse to collect credentials first and then use a phone call to bypass caution. Defenders reduce the risk by requiring call-back verification, limiting what can be approved by voice alone, training staff to distrust urgent requests, and using out-of-band checks for account recovery and payment changes.