Launch of the Groundbreaking KAS Study on AI in South African Newsrooms

The Konrad-Adenauer-Stiftung (KAS), through its Media Programme Sub-Saharan Africa, released a pivotal report on March 31, 2026, titled Navigating Risks and Rewards: How South African Journalists Use AI in the Newsroom. This mixed-methods study, conducted by the Centre for Information Integrity in Africa (CINIA) at Stellenbosch University and Karen Allen International (KAI), offers one of the first in-depth looks at artificial intelligence (AI) adoption among South African journalists. Launched on April 10, 2026, in Johannesburg, the event featured presentations by lead researcher Karen Allen and Prof. Herman Wasserman, alongside a panel including executives from the Press Council of South Africa and Arena Holdings. The discussion underscored AI's rapid integration into newsrooms, often without adequate safeguards.

Authors highlighted journalists' enthusiasm tempered by caution, with Prof. Wasserman noting, "South African journalists are navigating AI with both curiosity and caution. But without the necessary support systems in place, the risks may outweigh the benefits." This study arrives amid global AI advancements, addressing a critical gap in African contexts where resource constraints and linguistic diversity amplify unique challenges.

Methodology: Capturing Real-World Insights from SA Journalists

Employing surveys and in-depth interviews across print, broadcast, and digital media, the KAS study captures perceptions from journalists nationwide. While exact respondent numbers are not specified in the press release, the qualitative depth reveals patterns in AI use driven by individual initiative rather than top-down strategies. Complementary research, such as the Centre for News, Technology & Innovation (CNTI) survey of 1,012 South Africans and broader African journalists, shows 81.7% using AI tools, far exceeding global averages for tasks like writing enhancement (74% vs. 56%) and fact-checking (69% vs. 52%).

These findings align with a 2024 qualitative study of 17 South African journalists, who viewed AI adoption as inevitable, enhancing productivity while raising bias concerns. The State of the Newsroom 2025 report further corroborates cautious experimentation in commercial, independent, and community outlets.

Current Extent of AI Adoption Across South African Media

AI has permeated South African newsrooms unevenly but pervasively, with journalists embedding it informally for routine tasks. The KAS report notes widespread use without institutional frameworks, echoing global trends where 75% of Global South journalists apply AI in newsgathering, production, or distribution. In South Africa, adoption mirrors this, with Media24 appointing a Head of AI and forming a 'GenAI circle' for brainstorming, while Daily Maverick uses it for summaries and headlines.

Community media like Pondoland Times leverages AI for SEO and editing, and CNBC Africa deploys virtual anchors. Yet, KPMG's 2023 estimate warns 43% of writing tasks could automate, fueling job fears amid conservative uptake. Public optimism is high: 74% of South Africans are comfortable with AI for accuracy checks, per CNTI.

Popular AI Tools and Applications in Daily Workflows

South African journalists favor accessible tools like ChatGPT for drafting, Grammarly for editing, and transcription software such as Toqan AI. Briefly News' 'Editorial Eye' boosts output from 80-90 stories daily via proofreading and style checks. Common uses include research, summarization, interview transcription, headline generation, social media content, and data analysis—tasks saving time in under-resourced environments.

Per the KAS study, these streamline verification and investigations, but manual double-checks negate some gains due to 'hallucinations' (AI fabrications). CNTI data shows higher regional usage for document analysis (75%) and story drafting (54%).

Benefits: Boosting Efficiency and Expanding Capacity

AI empowers resource-strapped newsrooms by accelerating workflows, enabling deeper investigations. Journalists report time savings for human-centric tasks like rapport-building, with potential to amplify marginalized voices via data-driven insights. Media24's hackathons foster innovation, while AI aids multilingual translation, vital in linguistically diverse South Africa.

• Research and fact-checking: Faster background scans.
• Transcription: Automates hours of manual work.
• Content personalization: Tailors social media for engagement.
• Investigative support: Analyzes large datasets.

Karen Allen emphasizes, "Journalists have traditionally embraced innovation, but AI presents a mixed bag."

Major Challenges: From Training Gaps to Ethical Dilemmas

The KAS study flags informal adoption's pitfalls: scant training leaves journalists self-taught, vulnerable to ethical lapses. No formal policies heighten risks of plagiarism, copyright breaches, and bias amplification. Reuters Institute notes Global South journalists seek ethical guidance amid misinformation fears.

Over-reliance erodes skills, producing generic content lacking nuance. Public trust wanes with inaccuracies, as 81% of South Africans worry about misinformation—above global 59%.

Linguistic Limitations and Cultural Relevance Issues

AI falters with indigenous languages like isiZulu and isiXhosa, undermining accuracy in local reporting. The KAS report urges context-tailored tools, echoing Masakhane's African language AI efforts. This exacerbates inequalities, as Western-biased models marginalize township stories.

Panelist Lebogang Mokoena from Arena Holdings stressed adapting AI for South African realities during the launch.

Implications for Journalistic Standards and Public Trust

Without guidelines, AI risks homogenizing reporting and eroding the 'human soul' of journalism, per Press Council discussions. Transparency mandates, like Daily Maverick's disclosures, build trust. Yet, deepfakes and 'counterfeit journalists' threaten integrity, demanding Press Council updates.

Hendrik Sittig warns, "Without training... it also carries significant risks." Long-term, balanced adoption could fortify media resilience.

Higher Education's Role: Preparing Future Journalists for AI

South African journalism schools lag in AI curricula, with many offering data journalism but few dedicated modules. Stellenbosch's CINIA leads research, while Walter Sisulu University includes AI in 'Big Data, Journalism, and Society'. Universities like UJ, UCT, and Unisa provide AI short courses for educators, adaptable to media training.

Experts call for hands-on AI integration, partnering with computer science to bridge gaps. This aligns with KAS recommendations for educator support.

Real-World Case Studies from Leading Newsrooms

These examples illustrate strategic adoption balancing innovation and caution.

Photo by Markus Winkler on Unsplash

Recommendations and Path Forward

The KAS study advocates structured training, AI policies, ethical frameworks, and localized tools. Collaboration among newsrooms, academia, and funders like KAS is key. Press Council guidelines on disclosure could standardize practices, ensuring AI augments rather than supplants human judgment.

• Invest in multilingual AI development.
• Mandate transparency in AI use.
• Upskill via university partnerships.
• Monitor job impacts proactively.

By 2030, proactive steps could position South African journalism as AI-resilient, per expert outlooks.CNTI AI journalism report