Cloned - AI-powered tax scams target South Africans ahead of SARS deadline - IOL

South Africans’ growing trust in Artificial Intelligence (AI) is giving cyber criminals another weapon as taxpayers rush to meet the filing deadline, with experts warning that convincing fake South African Revenue Service (SARS) messages, investment offers, and tax documents are becoming harder to distinguish from the real thing.

This comes as the October 23 deadline approaches, with scammers exploiting the urgency around refunds, assessments, and outstanding tax matters to push people into acting before they stop to verify a message. 

Anna Collard, SVP Content Strategy & CISO advisor at KnowBe4 Africa, and Chris Norton, general manager for Sub-Saharan Africa at Kaspersky, warn that AI is making familiar fraud tactics more convincing. SARS has also warned taxpayers about scams and phishing attempts during the 2026 filing season.

Research by KPMG and the University of Melbourne found that 62% of South Africans surveyed were willing to trust AI, compared with a global average of 46%. KnowBe4 research also found that 86% of South African respondents agreed that deepfakes had become so realistic that it was harder to know what to trust.

Collard said scammers no longer needed to rely on obvious warning signs such as poor grammar, fake logos, or unnatural voices.

“So, people accept that the threat is real and hard to detect. They just don't believe it applies to them.

“We rate the risk accurately but rate ourselves generously. We think bad things happen to other people. And generative AI has removed the bad grammar, the wrong logo and the stiff voice, while leaving our confidence entirely intact,” she said.

For taxpayers, she said, a professional-looking message or one appearing to come from a familiar source cannot automatically be trusted.

“Trust should attach to the channel, not the content. A video of someone you recognise, a message in a familiar voice, a polished PDF, none of it is evidence of anything anymore,” she added.

Collard also warned against relying on AI chatbots for important financial decisions. 

“AI tools are fluent, and we mistake fluency for expertise. Use these tools to understand concepts and formulate better questions. Don't use them as your adviser.” 

Norton said the approaching tax deadline was creating another opportunity for scammers. Kaspersky researchers identified phishing websites designed to imitate SARS eFiling, as well as malicious attachments carrying names such as “Tax Returns Of RXX,XXX.XXX” and “Final_Assessment_Notification-2026-0X-0X.pdf”.

He said fake emails often carry phrases such as “Refund due” and “SARS Letter of demand” to persuade recipients to open attachments or follow instructions.

“By this point in the tax season, people are used to receiving messages about returns, assessments, and refunds, so an email like this may not immediately look out of place.

“That is what scammers rely on. The closer we get to the deadline, the easier it is to act quickly because you think you are dealing with something that needs your attention,” Norton said.

SARS has also warned that scammers are using AI to create professional-looking email templates and fake refund notifications. The revenue service said it will never request passwords, one-time PINs, banking PINs, or eFiling login credentials through email, SMS, social media, or telephone.

Norton warned taxpayers to also be careful about what information they feed into AI services when dealing with tax matters.

“AI can make complicated information easier to work through, but taxpayers should be careful about the information they provide to an external service. A tax return contains a detailed picture of someone’s financial and personal life. Sensitive records should stay out of public AI tools, and important tax information should be checked through trusted sources,” Norton said.

As the deadline approaches, taxpayers are advised to access SARS eFiling and other official digital channels directly to check assessments, refunds, and tax information rather than clicking links in unexpected messages.