🇳🇬 English, 6 October 2026, AI security risk
Half of South Africans Bypass AI Tool Restrictions
Employees are using unapproved AI tools, exposing companies to new security risks.
The use of unapproved AI tools by employees is a growing problem for companies, with 50% of South Africans surveyed having bypassed enterprise AI restrictions over the past year. This creates a new kind of threat perimeter, where sensitive information can leave a company’s controlled environment without a conventional cyberattack. Managers are more likely to be guilty of this than their teams, and unauthorised AI use is a bigger data-security problem than companies assume.