AI-powered scams are rapidly reshaping the fraud landscape, exposing new vulnerabilities within financial systems and eroding public trust in digital communication.
In recent months there has been a notable increase in cyberattacks that exploit AI to deceive individuals, often by imitating real people with alarming accuracy. A key strategy involves leveraging AI to produce highly convincing fake images, videos and audio, commonly referred to as deepfakes. They are used to impersonate real individuals and spread misleading or false information.
While early examples of deepfakes were often unconvincing, recent advancements have made them increasingly difficult to detect, making it easier for bad actors to mislead, manipulate and defraud.
Momentum Group’s financial director Risto Ketola recently disclosed he had been impersonated on WhatsApp. Cybercriminals used his LinkedIn profile photo to create a closed WhatsApp group, falsely presenting themselves as Ketola. While the incident did not involve AI-generated imagery or video, it highlights the significant harm that can result when an individual’s likeness is misused for malicious purposes.
Financial institutions must urgently evolve their fraud prevention strategies to stay ahead of sophisticated threats
Deepfake-driven cybercrime has escalated to the point where the South African Banking Risk Information Centre (Sabric) recently issued a strong warning about the growing threat of AI-enabled fraud.
Sabric specifically highlighted the use of deepfakes and voice cloning to impersonate bank officials, promote fake investment schemes and fabricate endorsements from well-known public figures.
The emerging threat not only compromises the integrity of the financial sector but also erodes customer trust and confidence in digital interactions.
In practice, fraudsters are increasingly using AI to bypass security measures such as automated onboarding systems and know your customer checks, enabling them to create accounts and access services under false identities.
From a business email compromise standpoint, attackers are incorporating deepfake audio and video of senior executives into phishing attempts, convincing employees to release funds or disclose sensitive information.
Social engineering attacks have also become more sophisticated, with AI used to analyse and replicate communication styles based on publicly available information, making scams appear more authentic.
In some cases, AI is used to generate entirely synthetic identities, combining real and fabricated data to create fake personas capable of applying for credit, laundering money or committing large-scale financial fraud.
Unfortunately, many legacy fraud detection tools aren’t designed to detect fake audio and video, making deepfake scams harder to spot.
In response, financial institutions must urgently evolve their fraud prevention strategies to stay ahead of sophisticated threats.
Regulators expect financial institutions to keep up with the latest cybercrime trends and failing to detect deepfake-based fraud can result in compliance failures, fines and legal action. Furthermore, financial institutions must consider the broader impact of the risks on customer trust.
As awareness of deepfake threats grows, it is understandable clients may begin to question the authenticity of video calls, digital signatures and other remote interactions. The erosion of confidence has the potential to hinder digital transformation initiatives and may even prompt some customers to disengage from digital platforms altogether.
• Sameer Kumandan is MD of SearchWorks
For opinion and analysis consideration, e-mail Opinions@timeslive.co.za
TimesLIVE
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.