World Bank AI Policy Audit Exposes South Africa's Missing Strategy Framework
Fabricated sources force South Africa to rebuild AI policy framework from scratch
SOUTH AFRICA FAILS TO REGISTER ON WORLD BANK’S AI POLICY MAP
South Africa’s name is absent from the World Bank’s map of nations with functioning artificial intelligence strategies, a gap that reflects a concrete operational failure rather than a paperwork delay. Released this week, the World Development Report 2026: The Promise of Artificial Intelligence documents more than 80 countries that had published national AI strategies as of June 2026. South Africa is not among them.
A footnote in the report states the reason plainly: South Africa “is not included as an economy with a strategy” because the country withdrew its draft national AI policy in April 2026. For a nation positioned as the continent’s most technologically advanced economy, the exclusion marks a serious setback.
The withdrawal traced back to a fundamental credibility failure. Communications minister Solly Malatsi pulled the 86-page draft on 26 April after internal verification revealed that its reference list contained fabricated sources. The document, gazetted on 10 April with 67 citations, included references to academic journal articles that did not exist. Most of those citations appear to have been generated by an AI tool and released without proper verification. Malatsi characterized the failure as more than a technical oversight, stating it had compromised the policy’s integrity at its core.
The World Bank defines a national AI strategy as a standalone, government-led framework articulating a country’s vision and action plan for artificial intelligence adoption. Its analysis of current global positioning reveals a stark disparity in preparedness. Among 25 low-income countries surveyed, only Rwanda has published a dedicated AI strategy. More than half of high-income economies have done so. South Africa, classified as an upper-middle-income nation, now stands outside both categories entirely.
Following the withdrawal, Minister Malatsi appointed an independent expert panel chaired by Wits University AI researcher Prof Benjamin Rosman to reconstruct the policy from the ground up. The rebuilding process has extended the timeline considerably. The department of communications and digital technologies has informed parliament that the panel’s consolidated report is expected in August 2026, with a revised draft scheduled for cabinet submission in November. Public consultation is targeted for January 2027.
Those timelines mean South Africa will not have a published national AI policy before next year, nearly three years after the National AI Summit that initiated the process in April 2024.
The situation carries an additional layer of awkwardness. Both the department of communications and digital technologies and the department of science, technology and innovation appear in the World Bank report’s acknowledgements as organizations that provided input to the research. The communications regulator Icasa and the South African Revenue Service are similarly credited as contributors. The country they represent, though, remains unrecognized on the report’s map of functioning AI strategies.
The World Bank’s broader analysis identifies infrastructure as central to AI implementation capacity. Reliable electricity, broadband connectivity, and foundational literacy represent the complements that the private sector will not independently fund, the report argues. Without these elements, AI tools deliver minimal practical benefit. South Africa’s performance on these measures is uneven. The country faces particular challenges on electricity supply, which the World Bank highlights specifically in its chapter on AI and environmental impact. The report cites Brazil and Malaysia as emerging destinations for hyperscale data centre operations, pointing to South Africa’s chronic load shedding as an example of how grid constraints actively deter large-scale AI investment.
Whether the Rosman panel’s August report will produce a draft that survives cabinet scrutiny, and whether the January 2027 public consultation target holds, remains the practical question now facing the department.
Q&A
Why was South Africa's draft national AI policy withdrawn in April 2026?
Communications minister Solly Malatsi withdrew the 86-page draft after internal verification revealed that its reference list contained fabricated sources, including academic journal articles that did not exist. Most citations appear to have been generated by an AI tool and released without proper verification.
What is the new timeline for South Africa's AI policy development?
An independent expert panel chaired by Wits University AI researcher Prof Benjamin Rosman is reconstructing the policy. The panel's consolidated report is expected in August 2026, with a revised draft scheduled for cabinet submission in November 2026 and public consultation targeted for January 2027.
How does South Africa's absence from the World Bank's AI strategy map compare to other nations?
The World Bank's World Development Report 2026 documents more than 80 countries with published national AI strategies as of June 2026. Among 25 low-income countries surveyed, only Rwanda has published a dedicated AI strategy. More than half of high-income economies have done so. South Africa, classified as an upper-middle-income nation, stands outside both categories.
What infrastructure challenges does the World Bank identify as constraining South Africa's AI implementation capacity?
The World Bank identifies reliable electricity, broadband connectivity and foundational literacy as critical complements for AI implementation. South Africa faces particular challenges on electricity supply; the report cites the country's chronic load shedding as an example of how grid constraints actively deter large-scale AI investment, with Brazil and Malaysia highlighted as preferred destinations for hyperscale data centre operations.