South Africa opens public debate on new AI policy as government moves toward regulation

A turning point for AI in South Africa

Artificial intelligence is already shaping how South Africans work, learn, and access services, but until now, it has largely operated without a formal national rulebook.

That is beginning to change.

The South African government has published its draft national AI policy for public comment, marking the start of a formal process to shape how artificial intelligence will be governed in the country.

Approved by Cabinet on 25 March, the draft policy is open for public input until 10 June.

Built on the Constitution and public interest

At the centre of the draft policy is a clear guiding principle: South Africa’s Constitution.

The Department of Communications and Digital Technologies, which led the policy development, says the framework is grounded in constitutional values, human rights protections, and the goal of inclusive economic growth.

While AI is already used across both government and private sectors, it currently operates under fragmented internal rules rather than a unified national law.

This draft policy aims to change that direction by creating a more structured national approach.

Not yet law, but already influential

Legal experts from Bowmans law firm describe the draft as an important early step in shaping South Africa’s AI landscape.

According to partners Robby Coelho and Justin Ramages, along with senior associate Ashleigh Brink and associate Tanya Chivauras, the policy is not binding law but it clearly signals where regulation is heading.

They note that the document should be seen as a starting point rather than a final legal framework, but it already outlines how government intends to approach AI governance.

Even without legal force, it already influences expectations around ethics, accountability, and transparency.

AI as a tool for growth and inclusion

The draft policy positions AI as a driver of economic inclusion, job creation, cost reduction, and development across Africa.

It identifies key sectors where AI could have the most impact, including:

• Education
• Healthcare
• Agriculture
• Public administration

At the same time, it emphasises the need to address South Africa’s inequality gap, digital divide, and broader socioeconomic challenges.

The policy also highlights that AI systems must be developed and deployed in ways that align with human rights standards and public interest.

New systems for oversight and accountability

One of the most significant elements in the draft is its proposed institutional structure.

It includes plans for multiple new bodies, such as:

• A National AI Commission or Office
• An AI Ethics Board
• An AI Regulatory Authority
• An AI Ombud’s Office
• A National AI Safety Institute
• An AI Insurance Superfund

Together, these institutions would form the backbone of future AI oversight in the country.

The policy also proposes a staged regulatory approach, starting with high-risk AI applications, before expanding to medium- and low-risk systems over time.

A phased and risk-based approach

Rather than introducing blanket rules, the draft suggests a risk-based model.

In practice, this means:

• Strict regulation for “unacceptable-risk” AI uses
• Targeted rules for high-risk systems
• Gradual guidance for medium- and low-risk applications

Sector-specific working groups and implementation roadmaps are also part of the plan.

This approach reflects global trends in AI governance, where regulation is increasingly tailored to risk levels rather than treated as a one-size-fits-all system.

What it means for businesses and the public

Although the policy is still in draft form, legal experts say its publication already matters for companies working in AI.

It signals a stronger regulatory focus on:

• AI ethics
• Transparency
• Data protection
• Fairness
• Sector-specific oversight

Businesses operating in South Africa’s digital economy are now expected to prepare for a future where AI governance becomes more structured and enforceable.

Public participation now becomes key

With the policy now open for comment, South Africans have until 10 June to weigh in.

This stage is critical not just for government and industry, but for everyday users whose lives are increasingly shaped by AI systems, from online services to workplace tools.

What emerges next will likely set the tone for how South Africa balances innovation with rights, regulation, and inclusion in the age of artificial intelligence.