Are the machines rising? South African finance and security facing looming AI threats
As AI evolves from simple assistants to autonomous agents making complex real-world decisions, South African financial institutions face both massive operational opportunities and sophisticated new fraud threats.
Artificial intelligence is rapidly evolving, with the rise of agentic AI posing significant challenges for South Africa's finance and security sectors. Agentic AI refers to trust in AI systems to autonomously make critical decisions, far beyond simple tasks like drafting emails. At a recent roundtable discussion in South Africa, experts explored how this technology is transforming banking, payments, and business workflows.
Mick Amelishko, AI Advocate at Sumsub, emphasized that the key shift is the agency granted to AI. While AI was once merely a helpful assistant, it now has the ability to interact with the external world, make independent transactions, and even register on websites or access bank accounts with permission. This rapid shift from human-centric transactions to AI-driven actions raises concerns about security and accountability.
Kyle Reese's famous warning from the 1984 film "The Terminator" highlights the potential for autonomous digital threats, though without the physical violence. In South Africa, enterprise banking trends show limited adoption of true AI agents compared to global markets. Advanced digital banking and instant payment systems make consumers comfortable with digital onboarding, but full agentic adoption is still early.
Major banks like Discovery Bank are pushing into agentic territory, yet South Africa lags behind countries like China. A key challenge is determining accountability when AI systems make mistakes or engage in fraudulent activities. Experts agreed that AI itself cannot be legally blamed, but humans who issue prompts or control systems bear ultimate responsibility.
Despite regulatory frameworks struggling to keep pace with technological advancements, fraud remains a growing concern. Sumsub data shows that while global identity fraud rates slightly decreased in 2025, highly sophisticated fraud attacks surged by 180% year-over-year, with AI-driven attacks becoming more prevalent. Balancing the convenience of agentic AI with vigilant human oversight will be crucial to maintaining security and trust in South Africa's evolving financial landscape.
Written by urgent.news from IOL's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
