SARS Warns Taxpayers About AI-Generated Refund Phishing

The South African Revenue Service (SARS) warned during the 2026 filing season that scammers are using AI to generate convincing refund-themed emails and SMS messages, with lures claiming refunds of up to R50,000 and directing taxpayers to fraudulent websites. BusinessTech reported the warning on July 29, noting that AI-generated templates make the messages harder to recognize as fraudulent. SARS advises taxpayers to verify notices through official channels and never disclose passwords, PINs or banking credentials.

SARS Warns Taxpayers About AI-Generated Refund Phishing SARS warned during South Africa's 2026 filing season that scammers are using AI to make fake refund emails and SMS messages look more convincing. BusinessTech reported on July 29 that the lures can claim refunds of up to R50,000 and direct taxpayers to fraudulent sites. SARS tells taxpayers to verify notices through official channels and never disclose passwords, PINs or banking credentials. The South African Revenue Service SARS warned during the 2026 filing season that scammers are using AI to make refund-themed emails and SMS messages look more convincing. BusinessTech reported the warning on July 29 and said the latest lures claimed refunds of up to R50,000 before directing recipients to fraudulent websites. BusinessTech quoted SARS as saying scammers were using AI to generate professional-looking email templates that are harder to recognize as fraudulent. That AI assessment comes from the publication's report of SARS's warning; the public SARS scam page documents the current examples and safety guidance but does not publish a technical analysis of how the messages were produced. Current refund lures SARS's official Scams and Phishing page lists a "Tax Return Notification - Action required" example dated July 28 and an "approved" tax-return email dated July 22. The agency says fraudulent messages can imitate SARS correspondence and link to fake forms or websites intended to collect personal and banking information. Accounting Weekly's July 28 account of the July 22 alert describes an email using a real SARS employee's name, a precise R68,652.86 refund amount and a PDF attachment containing a phishing link. That was a separate documented variant of the same refund-impersonation pattern, not evidence that every message uses the same amount or attachment. The timing matters because taxpayers who recently filed returns may reasonably expect assessment or refund notices. A polished message, a plausible amount and familiar branding can therefore create urgency even when the sender or destination is fraudulent. What recipients should verify SARS says it will not request passwords, one-time PINs, banking PINs or eFiling credentials through email, SMS, social media or telephone. It advises taxpayers to avoid links and attachments from unknown sources and to verify tax information by independently opening official SARS channels. For security and finance teams, the broader lesson is that grammar and visual polish are weak phishing signals once templates can be generated or refined cheaply. Sender domains, link destinations, attachment behavior and independent portal checks provide stronger evidence. BusinessTech reported that SARS asked recipients to delete and block scam notices and send uncertain messages to its IT security team. The reported campaign is a current phishing warning, not evidence of a breach of SARS systems. Its practical risk is impersonation: a convincing notice can steer a taxpayer away from an official channel and into a credential-harvesting workflow. Key Points - 1BusinessTech reported on July 29 that SARS had seen scammers use AI to create convincing refund-themed email and SMS templates. - 2SARS's official scam page documents current July refund lures and says it will not request passwords, one-time PINs, banking PINs or eFiling credentials through unsolicited messages. - 3The warning concerns impersonation and phishing, not a reported breach of SARS systems. Scoring Rationale The warning documents a practical use of generative AI in a financially sensitive tax-season phishing campaign. Its reach is regional, but the verification lessons apply broadly to identity, email-security and finance workflows. Sources Primary source and supporting public references used for this report. Practice interview problems based on real data 1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with. Try 250 free problems /problems