Understanding the Surge in AI Use Across South African Media

South African newsrooms are embracing artificial intelligence (AI) at a surprising pace, transforming daily workflows from research to content creation. A recent comprehensive study highlights how tools like chatbots and automated transcription are becoming staples, yet this rapid shift comes with significant hurdles. Journalists report using AI for tasks such as summarizing documents, drafting headlines, and even social media posts, often on an individual basis rather than through structured organizational strategies.

This integration reflects a broader global trend where AI augments human capabilities in resource-strapped environments, but in South Africa, it underscores unique pressures like multilingual demands and legacy infrastructure in media houses. The study's mixed-methods approach—combining surveys and in-depth interviews—paints a picture of enthusiasm tempered by caution, with AI seen as a productivity booster amid shrinking newsroom budgets.

Key Tools and Applications Driving Newsroom Efficiency

Common AI applications include generative tools for quick research, real-time transcription during interviews, and translation services to bridge South Africa's 11 official languages. Journalists at outlets spanning print, broadcast, and digital platforms described AI as a 'timesaver' for repetitive tasks, allowing more focus on core storytelling. For instance, one respondent noted using AI to process raw audio from press conferences, cutting hours off manual work.

However, adoption remains uneven. Larger media groups experiment with advanced features like data analysis for investigative pieces, while smaller newsrooms stick to free tools due to cost barriers. This patchwork approach boosts short-term output but risks inconsistencies in quality and ethics.

The Training Deficit: Self-Taught Journalists Navigate Uncharted Territory

A glaring gap is formal training. Most journalists learn AI through trial-and-error, YouTube tutorials, or peer sharing, with few newsrooms offering structured programs. This ad-hoc method leads to underutilization and heightened error risks, as users grapple with prompt engineering and output verification without guidance.

Experts like Prof. Herman Wasserman from Stellenbosch University emphasize the need for 'on-the-ground training' to harness AI responsibly. Without it, South African media risks falling behind global peers where upskilling is institutionalized. Universities, particularly journalism schools, are pivotal here, yet many curricula lag, incorporating AI only peripherally in data journalism modules.

Policy Vacuum Raises Ethical Red Flags

No formal AI policies exist in the majority of newsrooms, leaving journalists to self-regulate amid concerns over hallucinations—AI-generated inaccuracies—and plagiarism. Manual fact-checking negates efficiency gains, while fears of bias amplification loom large, given training data often skewed toward Western contexts.

The South African National Editors’ Forum (SANEF) and Press Council are drafting guidelines stressing disclosure and accountability, but implementation varies. A 2025 CNTI survey found only 13% of Global South newsrooms have policies, mirroring SA trends. This void erodes public trust, especially as audiences demand transparency on AI's role in stories.

Language and Context: The Need for Localized AI Solutions

AI struggles with indigenous languages like isiZulu and isiXhosa, producing subpar translations or summaries that lack cultural nuance. Journalists manually edit outputs, highlighting the scarcity of Africa-centric models. Calls grow for collaborative development of tools attuned to local dialects and storytelling traditions.

Read the full CINIA/KAS study for detailed insights into these linguistic challenges.

Photo by Sincerely Media on Unsplash

University Research Leading the Discourse

South African academics are at the forefront. Soligah Solomons and Musawenkosi Ndlovu from the University of Cape Town's qualitative study of 17 journalists revealed perceptions of AI as 'inevitable' yet bias-prone, urging ethical frameworks. Similarly, Stellenbosch's Wasserman warns of risks outweighing benefits sans support systems.

Access the UCT study here.

Journalism Schools: Bridging the Curriculum Gap

Southern African J-Schools like University of Johannesburg, Tshwane University of Technology, and Walter Sisulu University weave AI into data modules but lack dedicated courses. A 2025 study flags educator knowledge gaps and resource shortages, with students often outpacing faculty. Postgraduate research on AI ethics is burgeoning, signaling future reforms.

Read more on J-School gaps in this analysis.

Real-World Examples: AI in Action at SA Outlets

• Daily Maverick: Uses AI for transcription and sentiment analysis in investigations.
• eNCA: Experiments with automated alerts and social content optimization.
• News24: Leverages AI for headline testing and audience personalization.

These cases show productivity spikes but underscore verification needs to combat errors.

Risks and Broader Implications for Media Integrity

Unchecked AI could homogenize reporting, sidelining diverse voices and deepening inequalities. Public surveys indicate optimism (74% see tech positively impacting information access) but demand transparency. For higher education, this signals urgent curriculum updates to produce AI-literate graduates.

Path Forward: Training, Policies, and Collaboration

Recommendations include newsroom policies mandating disclosure, university-led training hubs, and public-private partnerships for local AI. SANEF's guidelines offer a start, with universities like UCT piloting workshops.

Photo by Sincerely Media on Unsplash

Future Outlook: AI as Ally in South African Journalism

With proactive measures, AI could empower under-resourced newsrooms, enhancing investigations and inclusivity. Universities must lead by embedding AI ethics in programs, preparing the next generation for a hybrid future.