This month, the European Union activated the world’s first comprehensive artificial intelligence legislation, imposing strict transparency and accountability mandates on digital content. While the EU AI Act was drafted in Brussels, its immediate ripple effects are hitting media houses across the continent. For South African publishers—and the broader continental media ecosystem—this regulatory milestone serves as a critical warning. As the integration of automated technology accelerates, regulating AI in African newsrooms is no longer a futuristic debate; it is an urgent necessity to protect the public trust that separates professional journalism from the disinformation rampant on social media.
The European framework does not force media companies to abandon automation. Instead, it demands clear labelling of synthetic content and robust human oversight. For publishers grappling with shrinking budgets, integrating AI in African newsrooms is an attractive strategy for summarising data, translating text, and accelerating distribution. Yet, a recent study by the Centre for Information Integrity in Africa (CINIA) reveals a troubling reality: most South African outlets lack formal AI policies. Adoption remains highly fragmented, driven informally by tech-savvy individuals, with some journalists admitting to using automated tools in secret out of fear or uncertainty.
This informal, unstructured approach poses severe editorial risks. Senior editors warn of creeping plagiarism, diminished investigative rigor, and the stark inability of Western-developed algorithms to navigate complex African languages. Because popular AI models are trained predominantly on datasets from the Global North, they frequently mistranslate nuanced cultural and politics terminology. In response, local developers are racing to build large language models tailored specifically to African dialects, ensuring that automated tools reflect the continent’s realities rather than distorting them.
The stakes for getting this right extend far beyond media ethics. When advanced technology is deployed without rigorous guardrails, the consequences can be catastrophic for institutional credibility. Recently, South Africa’s Communications and Digital Technologies Department was caught using fake, AI-generated citations in its draft National AI Policy. Even more alarming, a Johannesburg acting judge issued a ruling heavily reliant on hallucinated, non-existent case law. These high-profile failures highlight exactly why the Press Council of South Africa (PCSA) is pushing for stringent editorial oversight.
According to PCSA Executive Director Phathiswa Magopeni, the industry debate fundamentally revolves around verification, disclosure, and explainability. While South Africa’s vibrant fact-checking community has largely kept political deepfakes at bay, traditional broadcasters and print media must hold themselves to a much higher standard. As AI becomes deeply entrenched in research and production workflows, publishers must be entirely transparent with their audiences about where and how automated systems are used.
Moving forward, African media leaders cannot simply cut and paste European regulations, nor can they ignore the technological shift. The creation of a binding, Africa-centric AI press code is a survival imperative. As CINIA researcher Rejoice Malisa-van der Walt argues, regulations must be dynamic and inclusive, requiring active consultation with audiences to determine their tolerance thresholds for automated news. Ultimately, credibility remains the sole currency of professional journalism. If media organizations fail to establish clear, transparent rules of the road, they risk surrendering their vital democratic role to the unchecked chaos of the algorithm.
