South Africa: Prioritizes AI Regulatory Principles Over Strict Laws

• South Africa is prioritizing AI regulatory principles over immediate, strict rules due to rapid technological advancements.
• Financial institutions are already actively using AI for credit scoring and market analysis.
• A key concern is establishing clear liability for errors made by AI models, similar to existing gaps in corporate governance.
• Regulators are calling for a joint discussion paper on "agentic AI" and a clear mandate for human accountability in AI-driven decisions.
• The country is not seeking to directly copy foreign AI legislation like the EU AI Act.

This means ensuring that a human individual ultimately signs off on decisions that have been drafted or recommended by AI models, thereby retaining human accountability at the final stage of critical processes.

South Africa is currently adopting a cautious stance on formal artificial intelligence (AI) regulations, opting to prioritize the development of guiding principles rather than enacting immediate, strict statutory rules. This approach, championed by figures such as Financial Sector Conduct Authority (FSCA) Commissioner Unathi Kamlana, acknowledges the rapid evolution of AI technologies. The decision to focus on principles is not indicative of a pause in AI adoption within the financial sector; rather, it reflects an alignment with the ongoing integration of these advanced systems.

Financial institutions in the country are already extensively utilizing AI models for critical functions. These applications include sophisticated credit scoring mechanisms and comprehensive market scanning operations, demonstrating that AI is an active and integral component of current financial practices. The regulatory philosophy underpinning this strategy recognizes that prescriptive legislation often struggles to keep pace with the swift advancements characteristic of the AI landscape, making a principles-based framework a more adaptable solution for South Africa AI regulatory principles.

While a principles-based approach offers flexibility, it inherently creates significant challenges, particularly concerning accountability and liability when AI models produce erroneous outcomes. This regulatory gap mirrors existing issues observed in other legal frameworks, such as the Companies Act, where pinpointing responsibility for boardroom decisions can be complex. The absence of clear guidelines on who bears the ultimate burden when an AI system makes a mistake is a critical concern that needs urgent attention.

The rapid deployment of AI, including advanced "agentic AI" systems capable of autonomous decision-making, necessitates a robust framework for oversight. Without explicitly designated accountable officers, any set of governance principles risks becoming circular and ineffective. A comprehensive market scan has already identified numerous AI use cases within the South African context, further underscoring the immediate need to address these liability questions.

To effectively navigate the complexities of AI integration, a crucial next step involves the release of a joint discussion paper specifically addressing "agentic AI" and its implications. This document is expected to provide much-needed clarity and direction for the industry. It is widely understood that South Africa does not require an immediate replication of comprehensive foreign legislation, such as the European Union's AI Act, which might not be suitable for the local context.

Instead, the focus should be on establishing fundamental safeguards. A key demand from regulatory bodies, including the Reserve Bank, the Prudential Authority (PA), and the Financial Sector Conduct Authority (FSCA), is the explicit affirmation that human oversight remains paramount. This means ensuring that a human individual ultimately signs off on decisions that have been drafted or recommended by AI models, thereby retaining human accountability at the final stage of critical processes. International frameworks, such as those emerging from the G20, are unlikely to provide this specific, localized guidance for Pretoria.

Source: Original reporting via financial sector analysis.