Last updated: 25 August 2026
Mzansi Political Safari uses technology, including artificial intelligence and automation, to support parts of its newsroom and publishing workflow. This policy explains the limits we place on those tools and how responsibility for published material is maintained.
How AI may be used
AI tools may assist with tasks including:
- research planning and discovery of potentially relevant public sources;
- summarising source material for internal review;
- transcription and extraction of key points;
- drafting or restructuring copy;
- grammar, style and readability checks;
- headline, excerpt, taxonomy and SEO preparation;
- comparison of statements, timelines or public records;
- production workflows, formatting and administrative tasks.
AI is not a source
AI-generated output is not treated as evidence or as an independent source. Factual claims intended for publication should be checked against the underlying documents, direct statements, official data, reputable reporting or other appropriate sources.
Human editorial responsibility
A human editor or publisher retains responsibility for material published by Mzansi Political Safari. AI output may be changed, rejected or supplemented and should not be published merely because a tool generated it.
Accuracy and fabricated information
Generative AI can produce incorrect, outdated or fabricated information. Names, dates, quotations, statistics, legal claims, election information and other material facts require particular care. Direct quotations should be verified against the original source wherever reasonably possible.
Political neutrality and bias
AI systems may reproduce bias present in training data or prompts. AI-assisted political coverage remains subject to the same standards of neutrality, attribution, fairness and consistency as other Mzansi Political Safari reporting.
Disclosure
We maintain this public policy as a standing disclosure that AI may assist our production process. Where AI plays a substantial or unusual role in the creation of a specific published item, or where a reader could reasonably be misled without disclosure, we will consider an article-level disclosure describing that use.
Images, audio and video
AI-generated or materially AI-altered visuals, audio or video must not be presented as authentic documentary evidence of a real event. Where such media are used as illustrations, they should be clearly identified where needed to prevent misunderstanding.
Confidential and sensitive information
Unpublished sensitive material, personal information, confidential source information or legally restricted material should not be entered into an external AI service without appropriate consideration of confidentiality, data protection, security and the service provider’s terms.
Corrections
The use of AI does not reduce our responsibility to correct errors. An error introduced through an automated or AI-assisted process is handled under the same Corrections & Complaints Policy as any other editorial error.
Review
Because AI technology and newsroom practice evolve quickly, this policy will be reviewed periodically and updated when our material uses of AI change.
Contact
editorial@politicalsafari.co.za