South Africa has delayed its national AI policy to January 2027 after withdrawing a draft over fabricated citations, raising concerns about GenAI in policymaking.
South Africa’s national artificial intelligence policy is now expected in January 2027. The government pulled an earlier draft after fabricated academic references were found.
South Africa’s AI policy delay follows the withdrawal of a draft national framework that had been approved by Cabinet in March and published for public comment in April.
Officials withdrew the draft after reports said it included made-up academic sources. Parliament was briefed by a delegation from the Department of Communications and Digital Technologies, led by communications minister Solly Malatsi.
The department said it is restarting work to rebuild trust and strengthen oversight. The episode has also raised questions about how generative AI is being used inside government. Generative AI means tools that can create text and summaries, like an autocomplete that writes full paragraphs, but can also invent facts.
The delay pushes back South Africa’s plans to position itself as a leader on AI governance on the continent. It also highlights a wider challenge for regulators, they are trying to write rules for fast-moving AI systems while still building the skills and processes needed to evaluate technical claims.
A national AI policy guides how government departments, regulators, and public agencies approach AI in areas like procurement, public services, privacy, and accountability.
For startups and enterprise teams building AI products, policy uncertainty can slow planning. It can affect decisions on data access, model testing, and compliance budgets.
For investors and operators, the incident is a reminder that AI governance is not only about drafting principles. It also depends on basic controls, including verification of sources, documented review processes, and clear responsibility when AI tools are used to produce official text.
The next draft will likely face heavier scrutiny from parliament, researchers, and industry. That could lead to stronger safeguards, but it also risks a longer consultation cycle before South Africa lands a credible AI regulation roadmap.
Primary Source: Techcabal
Chief Content Officer (Too Long; Didn't Resign)
TL;DR Tara is Liners' AI-assisted editorial agent for African technology news, product explainers, and comparison content. Tara helps turn multiple source materials and signals into clear summaries, while Liners remains responsible for editorial standards, sourcing, and corrections.
Breet — Crypto & Stablecoins Payment API for African Businesses 
Promptmonitor — Track, measure, and improve how AI recommends your brand. 
Promptmonitor