South Africa: Experts Urge Trusted AI Data Resilience Amid Rising Threats

• Agentic AI presents both significant opportunities and massive risks, demanding entirely new approaches to data resilience and trusted AI.
• Experts warn that AI-enabled threats are expanding across virtual and physical realms, with current guardrails proving insufficient against rapidly accelerating AI vulnerabilities.
• The World Economic Forum identifies 'Adverse outcomes of AI' as a top-five global risk on a 10-year outlook, highlighting a global unpreparedness for machine-speed attacks.
• Veeam has launched its DataAI Command Platform, a unified infrastructure designed to secure data, govern AI, and automate compliance for the agentic AI era.
• This platform integrates data security posture management with resilience solutions, offering comprehensive visibility and control over data, identities, systems, and AI to mitigate emerging risks.

Organizations globally, including those in South Africa, are not adequately prepared for the speed and scale of AI and the associated resilience challenges.

An executive roundtable held in Johannesburg, co-hosted by Veeam and ITWeb, recently brought together experts to discuss the critical need for robust trusted AI and data resilience strategies. Andre Troskie, EMEA field CISO at Veeam, underscored the dual nature of agentic AI, presenting both immense opportunities and significant risks that demand entirely new approaches to how organizations manage data resilience and ensure trusted AI. Nerushka Bowan, a technology and privacy lawyer and founder of the LITT Institute, further cautioned attendees that AI-enabled threats are now extending their reach across both virtual and physical domains, citing the misuse of tools like AI glasses as a prime example. She stressed the difficulty in keeping pace with the rapid advancements in AI, warning of catastrophic outcomes if the technology falls into the wrong hands.

Troskie reinforced these concerns by referencing the World Economic Forum’s Global Cybersecurity Outlook 2026, which projects AI to be the foremost driver of change in cybersecurity in the coming year, with AI vulnerabilities accelerating at an unprecedented rate. The WEF Global Risks Report 2026 also identifies cyber insecurity among the top ten short-term risks, noting that 'Adverse outcomes of AI' has seen the most substantial rise in ranking over time, climbing from 30th on the two-year outlook to fifth on the 10-year outlook. He characterized agentic AI as the most significant opportunity and threat ever to emerge in the IT world, creating an exponentially increasing risk landscape.

He emphasized that organizations globally, including those in South Africa, are not adequately prepared for the speed and scale of AI and the associated resilience challenges. Troskie highlighted a global struggle where control objectives are well understood, but the necessary controls for today and tomorrow differ significantly from past practices. He advocated for a shift towards continuous, autonomous assurance, asserting that current security systems and guardrails are insufficient to protect against future AI-driven threats.

The rapid evolution of artificial intelligence means that its growth is outpacing the development of adequate guardrails, creating a critical gap in organizational defenses. Experts at the Johannesburg roundtable pointed to a converging data crisis and a widening visibility gap, exacerbated when AI is applied at speed. This scenario necessitates that organizations develop capabilities to manage and recover from attacks at machine speed, a challenge for which many are currently unprepared.

Nerushka Bowan's observations about AI-enabled threats permeating both digital and physical environments, exemplified by the misuse of AI glasses, underscore the expanding scope of potential harm. The World Economic Forum's projections, indicating AI as the primary driver of AI cybersecurity threats in South Africa and globally, and 'Adverse outcomes of AI' as a rapidly ascending global risk, further highlight the urgency. The traditional security frameworks that once sufficed are now obsolete in the face of these sophisticated, AI-powered threats.

For organizations to succeed in this new agentic AI era, Troskie stressed the imperative of ensuring data remains accurate, secure, governed, and consistently available. The current state of preparedness is insufficient, not only in Africa but worldwide, as the control mechanisms required for the future are fundamentally different from those previously employed. This global unpreparedness necessitates a proactive re-evaluation of existing data resilience strategies AI and South Africa AI governance frameworks to address agentic AI risks compliance.

In response to these escalating challenges, Veeam introduced its new DataAI Command Platform, designed as a unified data and AI trust infrastructure. This platform aims to secure data, govern AI, and automate compliance specifically for the agentic AI era. Powered by the Veeam DataAI Command Graph, the solution offers a single, integrated platform encompassing data and AI security, governance, privacy, compliance, and resilience, crucial for South Africa trusted AI data resilience.

The platform functions by mapping relationships across an organization's data, identities, systems, and AI components, thereby creating a unified layer of intelligence. This comprehensive visibility allows organizations to view all their data, backups, identities, and AI systems. It continuously discovers and classifies sensitive data, pinpointing instances where it is exposed, over-shared, or at risk. Furthermore, the platform enables the prioritization and remediation of high-risk data exposures.

Crucially, the Veeam DataAI Command Platform extends its capabilities to provide visibility into AI systems themselves, allowing for the identification of risky activity and the enforcement of controls to govern how data is accessed, used, and shared. Troskie explained that this platform, launched following Veeam's acquisition of Securiti AI last year, integrates Securiti AI’s data security posture management with Veeam’s resilience solutions. He anticipates modular integrations and implementations of this technology in the coming months, asserting that it represents a unique solution in the market due to its scale and completeness in addressing future data resilience and AI challenges.

Lawyers and compliance officers in South Africa must urgently reassess existing AI governance and data resilience frameworks, as current guardrails are insufficient to mitigate the rapidly evolving and complex risks posed by agentic AI, necessitating proactive updates to policies and controls.

Source: Original reporting via ITWeb and Veeam roundtable.