South Africa's ambitious endeavor to formulate a national Artificial Intelligence (AI) policy has encountered a significant setback, with Communications Minister Solly Malatsi announcing the withdrawal of the draft document. The policy, intended to guide the ethical and responsible deployment of AI, was withdrawn after an internal review uncovered numerous fabricated references, which are strongly suspected to have been generated by AI itself. Minister Malatsi underscored the gravity of the situation, stating, "This failure is not a mere technical issue but has compromised the integrity and credibility of the draft policy. As such, I am withdrawing the Draft National Artificial Intelligence Policy."

The irony of an AI policy being undone by AI's inherent flaws has not gone unnoticed. The draft, which aimed to establish national standards for AI across diverse sectors such as manufacturing, energy, and transport, appears to have fallen victim to "hallucination" – a phenomenon where AI confidently generates false information. What makes this incident particularly striking is that these false citations managed to pass through multiple stages of review, including approval by the Cabinet, led by President Cyril Ramaphosa, before being released for public consultation in April. The public was initially given until June 10, 2026, to provide feedback.

Malatsi acknowledged the high likelihood that AI-generated citations were incorporated without adequate verification, labeling it an "unacceptable lapse." He emphasized that this incident serves as a stark reminder of the indispensable role of human oversight in the age of artificial intelligence. The controversy quickly escalated into a political debate, drawing sharp criticism from Khusela Diko, chairperson of Parliament's communications committee. Diko publicly called for the policy's immediate scrapping, urging its resubmission "without using ChatGPT this time," a pointed remark that sparked a public exchange with fellow Cabinet member Dean Macpherson. This public spat highlighted underlying tensions within the Government of National Unity, which comprises members from various political parties serving under President Ramaphosa.

Before its unfortunate withdrawal, the draft policy was envisioned as a crucial component of South Africa’s strategy to navigate the Fourth Industrial Revolution (4IR). It proposed foundational principles, such as intergenerational equity, ensuring that AI innovation would benefit both present and future generations. Additionally, it aimed to delineate national priorities and devise sector-specific strategies for AI implementation. Deputy President Paul Mashatile had recently highlighted this policy as integral to the country's broader efforts to prepare for rapid technological transformation. Ultimately, this incident has ironically become a compelling case study illustrating the very risks and challenges that the withdrawn AI policy was designed to address: the critical need for meticulous verification when utilizing AI tools for drafting and research, especially concerning academic citations where accuracy is paramount to credibility.