South African Government Delays AI Policy After Reference Errors Exposed

The long-awaited national artificial intelligence (AI) policy in South Africa has been delayed until January 2027 following the discovery of fabricated academic references in an earlier draft. Communications Minister Solly Malatsi announced the revised timeline to parliament, acknowledging a “major credibility crisis” that undermined public trust.

A Setback for Continental Leadership

The initial policy framework, approved by Cabinet in March and publicly released in April, was withdrawn after media reports revealed several citations were either fictitious or referenced publications that did not exist. The incident has raised concerns about the use of generative AI in policymaking and highlighted weaknesses in government oversight.

Minister Malatsi stated that department officials failed to detect these issues before publication. Two employees have been suspended as a result of what he described as an embarrassing blow to South Africa’s reputation as a potential leader in African AI governance.

Strengthening Governance Amid Rapid Technological Change

The independent AI review panel, chaired by Professor Benjamin Rosman and including experts from various fields, will now rebuild the policy framework. The revised document is expected to undergo further scrutiny before being resubmitted to Cabinet later this year.

This delay underscores a broader challenge facing African governments as they attempt to regulate emerging technologies while simultaneously developing institutional capacity to understand them. South Africa’s experience serves as a cautionary tale about the risks of relying on AI for policymaking without robust verification processes.

With the revised framework not expected until 2027, South Africa will remain without formal national AI regulations even as businesses and government institutions increasingly adopt AI-powered systems.