X

AI practices and challenges for journalism in Africa

Laurence Dierickx

2025-01-18

Newsrooms on the African continent are at an interesting crossroads where structural challenges and opportunities intersect. The potential of AI technologies for journalism is significant, offering tools to improve news gathering, fact-checking and content production. However, its adoption and use faces several challenges specific to the African context, ranging from inadequate digital infrastructure to dependence on technologies from the « Global North ».

Artificial intelligence (AI) refers to computer systems or tools that can simulate intelligent human behaviour, such as data analysis, decision making and adaptation to different contexts. Every day, journalists around the world use AI technologies as part of their professional routines: search engines, social networks, voice assistants rely on AI technologies. Their practices also include automatic transcription of interviews, note-taking summaries, spell-checking, machine translation and writing assistance.

Journalists are likely to use three types of AI systems: predictive, which anticipates outcomes based on statistical and historical models; analytical, which analyses and interprets data to derive actionable insights; and generative, which manages content from training data. The rise of generative AI has fuelled the idea, often overhyped by marketing, that these technologies can transform the day-to-day practice of journalism. It’s fair to say that the AI market could reach €200 billion by 2025, reinforcing this trend.

AI technologies require training on large datasets to be effective. This training process is regularly outsourced to African countries, where working conditions are precarious and wages low. Kenyan or Ugandan workers have been paid between $1 and $2.50 per hour for tasks related to AI training, such as moderating and filtering content for Meta or Open AI. These tasks sometimes involved viewing violent content, including scenes of beheadings. In Madagascar, some French companies are also outsourcing this work, exploiting precarious wages while offering few opportunities for career advancement.

Adapting technology to African contexts

For newsrooms on the continent, AI offers many opportunities to automate repetitive tasks, personalise and customise content, enhance investigative capabilities, develop tools specific to African languages and contexts, or optimise limited resources. In practice, journalists’ interest is growing, but this enthusiasm is balanced by a sense of being under-equipped, whether in terms of training, skills, financial resources or appropriate ethical frameworks.

Generative AI seems to be a game changer, as access is no longer dependent on skills or budget. Journalists are discovering these tools in the same way as in Europe, with curiosity and trial and error. However, these technologies raise concerns about their ability to represent the realities of local communities and their ability to properly interpret cultural, social or political information specific to a country. In addition, the technologies are often generic and developed in the ‘global north’, meaning that they are not necessarily designed and adapted for the African continent.

AI technologies also present an opportunity for investigative journalism. In practice, the Dubawa project, led by the Centre for Journalism Innovation and Development (CJID) in Nigeria, launched a chatbot integrated with WhatsApp in May 2024, as well as an audio platform for monitoring and transcribing audio programmes. These tools are aimed at strengthening fact-checking efforts by enabling journalists to quickly verify information conveyed, especially on social networks. In 2022, the Nigerian media Dataphyte launched Nubia, an AI tool designed to turn datasets into alerts and reports.

In Kenya, media outlets such as The Nation and Royal Media are integrating AI technologies to analyse data, and their newsrooms have also benefited from specialised training. Training programmes have been organised for journalists from Kenya, Somalia and Zimbabwe, often by American and British universities or NGOs. A major problem, however, is that AI tools struggle to recognise the continent’s diverse languages and realities.

In Zimbabwe, Alice , a virtual presenter launched by CITE in September 2023, aims to modernise news delivery and find new ways to engage audiences. By harnessing the power of AI, Alice aims to meet the needs of a population seeking modern ways to access information, while introducing an interactive and personalised dimension. However, it has elicited mixed reactions from audiences (Ndlovu , 2024).

Between hopes and concerns

Some viewers initially believed Alice to be human, leading to frustration when she mispronounced some local names. There was also criticism of her lack of cultural and linguistic sensitivity, which was seen as detrimental to local identities and cultures. Her Western first name was also highlighted. Others hailed Alice as a sign of progress and innovation, seeing her as the future of news presentation and praising CITE for using AI to advance journalism. However, many in the audience expressed dissatisfaction with Alice’s inability to bring a « human touch » to her presentation, noting that her voice lacked emotional depth, making it difficult to connect with the content. Alice’s presentation also raised concerns about job losses for traditional journalists.

The introduction of AI in journalism in Africa raises a number of concerns shared by media professionals. According to a study by Mohammed et al. (2024), a number of ethical issues are at the centre of the debates. For example, 48% of respondents believe that AI raises ethical issues in journalism. This concern mainly relates to the impact of AI on the reliability, impartiality and transparency of information. In addition, nearly 30% of respondents believe that AI will have little or no impact on the financial stability of media organisations, suggesting that the economic benefits are still unclear. However, some claim that AI will improve journalistic practices (23%) and increase media productivity (18%).

An ally in the fight against disinformation

According to the Africa Center, disinformation campaigns aimed at manipulating African information systems have nearly quadrupled since 2022, with destabilising and anti-democratic consequences for African societies. At least 39 countries on the continent have been targeted by specific disinformation campaigns, and in many cases these countries have been attacked multiple times. The report, published in 2024, also targets Russia and China as major purveyors of disinformation. In the face of this threat, AI is emerging as a promising solution.

The African continent is home to several fact-checking organisations that play an important role in the fight against disinformation. One of the most prominent is Africa Check. Founded in 2012, it is the first independent fact-checking organisation in Africa, with offices in South Africa, Kenya, Nigeria, Senegal and the UK. PesaCheck is active in twelve countries in East and West Africa. Together with Africa Check and four other fact-checking organisations active on the African continent, Pesa Check is a member of the International Fact-Checking Network (IFCN). This membership strengthens the credibility and impact of these organisations in the fight against disinformation.

Whether in Africa or elsewhere in the world, fact-checkers are using tools based on AI technologies to cope with the growing volume of information to be verified and the speed at which fake news spreads. AI offers several advantages in the fact-checking process: identification and categorisation of information, content analysis to assess reliability, automation of verification tasks (images, data), detection of errors, inaccuracies, detection of manipulated images, detection of generated images, etc.

Some African initiatives specifically address the needs of fact checkers on the continent. These include MyAIFactChecker , a platform launched in 2024 by FactCheck Africa. It allows you to quickly check the authenticity of social media posts and other digital content. It has a user-friendly interface and supports a wide range of languages, including English, Swahili, French, Arabic and local Nigerian languages. The platform also offers a voice search function, making fact-checking more accessible.

According to Seck- Sarr (2023), although fact-checking initiatives in Africa show a strong desire to innovate, they remain largely dependent on funding and technology from GAFAM. Fact-checking programmes are also supported by funding from tech giants such as Google and Meta. While these partnerships are necessary for their operation, they reinforce the dependence of African initiatives on these large companies, which paradoxically contribute to the spread of the false information they seek to combat.

Many challenges to overcome

The introduction of AI into journalism in Africa has undeniable potential to transform the way news is collected, analysed and disseminated. However, its adoption is hampered by a number of structural and economic challenges. Key barriers include an unstable electricity supply, with frequent power outages disrupting the use of AI tools and making their application unreliable in many parts of the continent. Furthermore, the high cost of digital devices, including smartphones, limits access to AI technologies, especially in rural areas where infrastructure is inadequate. Furthermore, the digital divide between large cities, which are better equipped, and rural areas, which are often left behind, exacerbates inequalities in access to technological tools. Reliance on foreign solutions further exacerbates these inequalities in access to digital infrastructure, as big tech solutions are often inaccessible to a large proportion of the African population due to cost or limited connectivity.

In addition to these logistical challenges, Africa faces a critical lack of reliable and structured data needed to effectively train AI systems. Low levels of data collection, particularly in rural areas, and limited access to public data make it difficult to apply AI for in-depth analysis and high-quality research. The lack of robust systems to collect, store and share data, combined with insufficient financial resources to fund AI tools and related infrastructure, hampers the widespread adoption of these technologies.

These technological and infrastructural challenges are compounded by an inadequate legal framework, which further complicates the adoption and responsible use of AI in Africa. Current data protection laws are largely outdated and do not provide adequate protection against the risks inherent in the use of AI, including the mass collection and exploitation of sensitive data. In addition, cybersecurity remains a major concern, as the lack of adequate regulation leaves African countries vulnerable to cyberattacks, jeopardising the security of AI systems and the confidentiality of information. The legal framework is also incomplete when it comes to accessing and using public data, with laws still too vague and restrictive. This situation not only stifles innovation, but also limits the transparency needed for AI-based journalism.

In addition, the lack of research institutions specialising in AI in Africa prevents the development of local solutions adapted to the continent’s realities. The fear of seeing Africa digitally colonised by AI is very real. The countries of the continent are therefore faced with a dilemma: in order to preserve a form of digital sovereignty, they must master the use of AI, otherwise they risk becoming dependent on solutions imposed by foreign multinationals. This raises the crucial issue of technological emancipation, which requires the creation of local infrastructures and the training of appropriate skills.

But with 70% of its population under the age of 30, Africa is positioning itself as a key territory to harness the potential of AI. According to the African Union, AI could generate between $110 billion and $220 billion per year for African economies, while stimulating innovation and creating jobs. Several African countries have already put in place national AI development strategies, recognising its strategic and economic importance, and countries such as Mauritius, Egypt, South Africa and Tunisia are among the best prepared for the adoption of AI in Africa.

 

Bibliography

# # #