Malatsi and the Botched Draft AI Policy: What It Means for Agriculture in South Africa

April 30, 2026

In a surprising and embarrassing turn of events, Communications and Digital Technologies Minister Solly Malatsi withdrew South Africa’s draft National Artificial Intelligence Policy just weeks after it was gazetted for public comment. The reason? The 86-page document contained multiple fictitious academic citations — sources that do not exist — widely believed to be “hallucinations” generated by AI itself.

The irony was lost on no one: a national policy meant to guide responsible AI development in South Africa was itself compromised by the very technology it sought to regulate.

What Happened?

The draft policy was published on 10 April 2026. Within days, investigative reporting by News24 and others revealed that at least six of the 67 references in the document were fabricated. Some cited non-existent journals, while others attributed papers to real journals that had never published them. Minister Malatsi acted swiftly, withdrawing the draft on 26 April 2026, stating that the failure had “compromised the integrity and credibility of the draft policy.”

He described the lapse as unacceptable and promised consequences for those responsible, while committing to produce a revised version with proper verification processes.

Why This Matters for Agriculture

Agriculture was expected to be one of the key focus areas in the National AI Policy. The draft had recognised AI’s strategic importance for:

• Precision farming and resource optimisation

• Early pest and disease detection

• Climate-smart agriculture and drought forecasting

• Supply chain traceability and export compliance

• Smallholder farmer advisory services

The withdrawal creates a temporary policy vacuum at a critical time. South African agriculture is already undergoing rapid digital transformation, with tools like satellite monitoring (Farmonaut, Aerobotics), IoT sensors, generative AI agronomists, and drone spraying gaining traction. A clear, credible national AI framework was needed to provide:

• Regulatory certainty for AgriTech investors and startups

• Guidelines on data ownership and farmer privacy (especially important under POPIA)

• Ethical standards for AI use in food production and biosecurity

• Support for inclusive innovation that reaches smallholders and emerging black farmers

Immediate Implications for Farmers and AgriTech

1. Uncertainty Slows Investment

   Local and international AgriTech companies may pause major deployments until a credible policy framework is in place. This could delay rollout of new tools in 2026/27.

2. Missed Opportunity for Smallholders

   A well-crafted policy could have accelerated affordable, locally relevant AI tools (SMS/voice-based agronomists, satellite advisories, credit scoring). The delay risks leaving smallholders further behind.

3. Regulatory Risk Increases

   Without clear guidelines, issues around AI-generated advice liability, data sovereignty, and algorithmic bias in farm tools remain unresolved — creating hesitation among insurers, financiers, and exporters.

4. Positive Side: Opportunity for Better Policy

   The withdrawal allows time for genuine public consultation, broader stakeholder input (including farmers’ unions, Agri SA, TLU SA, and smallholder representatives), and proper verification processes. The next draft could be stronger and more practical for agriculture.\

The Path Forward

Minister Malatsi’s swift action to withdraw the flawed draft has been praised for showing accountability, even as it drew sharp criticism from opposition parties and the ANC. The real test will be how quickly and transparently a revised policy is developed.

For South African agriculture — already under pressure from climate change, water scarcity, biosecurity threats, and the need to close the smallholder-commercial productivity gap — a credible National AI Policy is essential. It should prioritise:

• Support for locally developed, low-bandwidth AI tools

• Protection of farmer data rights

• Incentives for AgriTech innovation that serves smallholders

• Clear ethical guidelines for AI use in food systems

Until a new policy is released, the sector will continue relying on existing regulations, industry self-regulation, and international best practices.

The botched draft was an embarrassing setback. But it also serves as a powerful reminder: when developing policy for transformative technologies like AI, especially in critical sectors like agriculture, human oversight, rigorous verification, and genuine consultation are non-negotiable.

Agriculture cannot afford another AI hallucination at the policy level.