“I learnt that while exploring these tools is exciting, it’s equally important to keep ethics in mind, making sure AI is used responsibly, fairly, and in ways that truly benefit people”. – Siposethu Matanga, Journalist, Pondoland Times
The rapid proliferation and expansion of AI in South Africa’s media sector presents a special opportunity for news outlets to improve their operations. Yet South Africa’s unique context – characterised by linguistic diversity and uneven digital access – means that an ethical and context-sensitive AI adoption approach is needed.
The feedback we gathered from the media ecosystem in South Africa also highlighted how a lack of AI policies and training opportunities was prompting journalists to take their own initiative, without support structures in place. These findings echoed insights summarised in the Foundation’s report – examining the use of AI in the Global South and emerging economies – where a majority of the 200 + journalists surveyed said that their newsroom had no established policy for AI, despite widespread usage. This increases the risk of AI-related missteps, such as unintentional bias or factual inaccuracies, which can impact reader trust and cause reputational damage. Ethical concerns were consistently raised in ecosystem feedback, reinforcing the need to support newsrooms in harnessing AI responsibly.
In response to this, and with support from Microsoft, the Foundation delivered a four-month programme to help four newsrooms in South Africa – The Mail & Guardian, AmaBhungane Centre for Investigative Journalism (amaB), Briefly News and Pondoland Times – leverage the potential of AI while minimising risk through establishing robust editorial policies and guardrails.
Developing AI-powered tools to address organisational needs
The participating newsrooms received bespoke mentoring by AI and journalism experts to develop AI adoption strategies and policies, a crucial step in laying the foundations for responsible AI adoption and innovation.
With the support of the mentors, the newsrooms identified practical applications for AI based on their organisational and business needs, and subsequently built AI tools that improved editorial efficiencies, enhanced audience reach, and streamlined processes. For example:
- Briefly News developed “Editorial Eye,” an AI-powered tool that assists with proofreading, style, and grammar checking. The tool has reduced manual copy editing, resulting in an increased output of 150-200 articles per day compared to 80-90 articles previously. This contributed to a 22% increase in page views over six months.
- AmaB developed a tool to package their in-depth investigations into more digestible and engaging multimedia formats, streamlining scriptwriting processes and making complex stories more accessible to new audiences.
- Pondoland Times developed a tool to auto-post content on social media and is currently working on enhancing its digital storytelling capabilities through an AI-generated avatar that will read the news flash. Their website traffic has increased from 2.4 million to 10 million monthly impressions for three consecutive months, correlating with a revenue increase of 20,000 ZAR in one month.
- Mail & Guardian developed a sub-editing tool that automated various processes, such as grammar checks, headline generation and hashtag suggestions, which has halved the amount of time previously spent on these tasks.

Deeper understanding of AI fundamentals and ethics
Throughout this process, newsrooms became empowered to use AI tools with confidence after initial reservations about potential job displacement and ethical breaches. The project reinforced their understanding of human oversight as a key component in ethical AI practices.
“Our primary goal was to ensure that what we were doing within the space was creating a tool that would assist our team rather than replace them, and ensure our readers continued to get content that was thought out from a human perspective.
The sub-editing [function] is basically marking your homework effectively. It’s used as a tool that is backstopped by human intervention the whole way along, and it requires human creative capacity and human creation to develop it”. – Douglas White, Circulation and Subscription Officer, Mail & Guardian
A shift in perception
This programme also facilitated close collaboration between participants, as they had the opportunity to learn from others in the industry.
“It was also an amazing opportunity to meet people from other newsrooms and group share in a safe and conducive space”. – Rianette Cluley, Director and Media Project Manager, Briefly News
“I think for too long we’ve viewed each other as adversaries in this space. And when you come together in a place like this, you realise that it’s not adversarial. It’s collaborative”. – Douglas White, Circulation and Subscription Officer, Mail & Guardian
Marking an important change in attitude and perception, the collaboration between participants and mentors fostered an environment geared towards peer learning that was based on trust and respect – crucial to the programme’s success.
More News
View All
World’s largest dataset shows transparency gaps in AI adoption
The Thomson Reuters Foundation…
Read More
The authoritarian playbook in action: Insights from Trust Conference 2025
Learn our Acting…
Read More
Antonio Zappulla: Technology is redefining power, information and influence. What is at stake?
View our CEO Antonio Zappulla’s opening remarks for Day Two of Trust Conference…
Read More
Antonio Zappulla: Three key drivers are reshaping the world and eroding democracy
View our…
Read More
Trust Conference 2025: Disinformation, lawfare and aid cuts — Civil society’s fight for survival
Trust Conference, the annual flagship forum hosted by the Thomson Reuters Foundation,…
Read More
Professor Can Yeğinsu joins Thomson Reuters Foundation Board of Trustees
We are delighted to…
Read More
Top priorities for Just Transition in Bangladesh
In one of the world’s most climate-vulnerable countries, a just…
Read More
AI Company Data Initiative drives transparency on corporate AI adoption
Now open for company…
Read More
Winners of the 2025 TrustLaw Awards Announced
This week marks the announcement of the 2025 TrustLaw Awards winners,…
Read More
How the Foundation is supporting LGBTQ+ communities worldwide
Over the last year, the Foundation has…
Read More