Accra, Ghana, November 3, 2026: Artificial intelligence is entering African newsrooms at a time when many journalists are struggling to survive on their earnings, working without formal contracts and facing threats for doing their jobs.
That is where the debate about responsible AI in African journalism should begin.
The technology offers practical opportunities. Journalists can use it to transcribe interviews, translate documents, organise large volumes of information, analyse datasets and assist with research. For smaller newsrooms with limited resources, some of these applications could make reporting more efficient and open up work that would otherwise be difficult to undertake.
But the arrival of new tools does not automatically improve the conditions in which journalism is produced. Nor does it guarantee that the benefits will reach the people doing the work.
A newsroom that cannot pay its journalists regularly, protect confidential sources or provide a safe working environment has problems that an AI subscription cannot solve.
The Federation of African Journalists (FAJ) brought this tension into focus during a two-day dialogue in Addis Ababa, Ethiopia, on 24 and 25 October 2026. Under the theme Advancing Responsible Artificial Intelligence and Decent Work for Quality Journalism in Africa, the meeting brought together journalist leaders and representatives of the African Union Commission, UNESCO, the International Labour Organization (ILO) and the European Union.
Participants examined the opportunities AI offers for research, translation and verification, alongside the risks of fabricated content, impersonation, disinformation, compromised editorial independence and job insecurity. They adopted a continental advocacy action plan intended to advance responsible AI, decent work and quality journalism.
The meeting's message was timely. But the evidence in a separate FAJ report suggests that the profession has a much more immediate problem to confront.
The crisis existed before AI entered the newsroom
FAJ's Report from a Continental Survey on Journalists' Working Conditions, Labour Rights, Safety and Media Freedom in Africa was published in October 2026. It gathered responses from 1,586 journalists in 45 African countries across the continent's five regions.
The figures on income and employment are stark. Sixty-one per cent of respondents said their income from journalism did not meet their basic living needs. Only four per cent said their salaries adequately covered essential costs. Fifty-eight per cent reported delayed salaries, 62 per cent worked without written contracts and 71 per cent lacked social protection.
These figures describe more than an employment problem. They raise questions about the conditions under which public-interest journalism can be sustained.
A journalist who does not know when the next salary will arrive may struggle to meet basic expenses. A freelancer without a written agreement may have little protection when an assignment is cancelled or payment is withheld. A reporter without health insurance or pension coverage may be expected to take professional risks without the security available to workers in other sectors.
The survey also found that freelancers accounted for 29 per cent of respondents, while contributors and stringers represented another 18 per cent. Full-time staff journalists made up 38 per cent.
This composition matters. A significant share of the people represented in the survey work outside conventional permanent newsroom arrangements. Yet they may still be expected to produce stories, meet deadlines, attend assignments and use new technologies to meet editorial demands.
The report is not a statistically representative census of all African journalists. Participation was voluntary, and responses were gathered mainly through unions, associations, professional networks and online platforms. The findings should therefore be read as an indicative account of experiences and patterns, not as precise estimates for every country.
Even with that qualification, the scale and range of the responses make the report an important contribution to the debate about the future of journalism.
Who will benefit from newsroom automation?
The ILO's February 2025 research brief, Generative AI and the media and culture industry, provides a useful framework for understanding the employment implications.
The researchers identify journalism, writing and translation among occupations with high exposure to generative AI. Exposure, however, does not mean that every job in those occupations will disappear. Some tasks may be automated, others may be supported by AI, and some roles may change as workers take on more responsibility for checking and supervising machine-generated material.
The ILO warns that the technology could also lead to job displacement, revenue losses and reduced demand for traditional roles. It argues for policy frameworks, fair remuneration, creative control and social dialogue between employers and workers.
That distinction between changing tasks and eliminating jobs is essential. It would be misleading to claim that AI will inevitably replace African journalists. It would be equally irresponsible to dismiss the possibility that media owners could use automation to reduce staffing costs or increase output without improving working conditions.
The question is who gets to decide how the technology is used and who benefits from the savings.
If AI helps a reporter process documents more quickly, the time saved could be invested in field reporting, interviews and verification. If it helps a small newsroom translate material into local languages, more communities could gain access to information.
But if the same technology is used to reduce the number of journalists while increasing the workload of those who remain, the gains may accrue mainly to employers. Journalists could be required to produce more content in less time, with fewer colleagues and no corresponding improvement in pay.
There is also the risk of replacing paid editorial work with automated summaries, rewrites and generic content. Such a shift could weaken the reporting that gives journalism its public value: original interviews, local knowledge, investigative work, scrutiny of public institutions and the ability to understand context.
This is not an argument against efficiency. It is an argument against treating reduced labour costs as the only measure of technological progress.
Media employers should explain how AI adoption will affect staffing, job descriptions, workloads and remuneration. Journalists should be consulted before systems are introduced, not informed after decisions have already been made. Training should be provided during working hours, and workers should have a say in how the technology changes their responsibilities.
The cost of poor working conditions is editorial independence
Perhaps the most consequential finding in the FAJ report concerns self-censorship.
Eighty-four per cent of respondents said they practised self-censorship frequently or occasionally because of intimidation or economic insecurity. Sixty-nine per cent said threats had led them to avoid certain stories.
The report also records political interference, barriers to accessing information, censorship and legal intimidation among the violations journalists experienced.
These findings point to a relationship between employment conditions and the freedom to report. A journalist who depends on irregular payments, fears dismissal or faces pressure from an advertiser may find it harder to challenge an employer's preferred narrative or pursue a politically sensitive investigation.
The survey does not establish that poor pay alone causes self-censorship. The pressures it documents include political interference, threats, restrictive laws and economic insecurity. But it makes clear that the ability to report freely cannot be separated from the conditions in which journalists work.
This is where AI governance needs to go beyond rules about checking machine-generated text.
A newsroom can introduce a policy requiring human oversight of AI-generated material and still fail to protect editorial independence if journalists are not free to challenge management decisions. It can demand accuracy while imposing unrealistic production targets. It can invest in digital tools while leaving reporters vulnerable to political pressure or retaliation.
Responsible AI requires responsible newsroom management.
The Center for News, Technology & Innovation's February 2026 briefing on newsroom AI policies reviewed 30 recent research papers. It found that existing policies commonly emphasise transparency, human supervision and verification, but often provide inadequate guidance on subtle biases embedded in third-party tools.
This is a practical warning. A newsroom policy cannot simply state that AI must be used ethically. It must explain what staff can and cannot do, which tools are approved, how information is checked, how confidential sources are protected and who is accountable when errors are published.
Human oversight must also be meaningful. Editors and reporters need the time, skills and authority to challenge an AI output. A requirement to check machine-generated material is not a safeguard if staff are under pressure to publish at speed without adequate resources.
Women journalists face a second layer of risk
The debate also needs a stronger gender perspective.
FAJ's survey found that 76 per cent of respondents believed women journalists faced additional risks in the media sector. The challenges identified included sexual harassment, online gender-based violence, unequal pay, workplace discrimination, exclusion from leadership and intimidation connected to political or investigative reporting.
The report also found that only 24 per cent of respondents confirmed that their workplaces had gender equality and sexual harassment policies. Forty-six per cent said their workplaces had no such policies, while 30 per cent did not know whether policies existed.
That last figure is important. A policy that staff do not know about, cannot access or do not trust is unlikely to provide meaningful protection.
AI adds another dimension to an already difficult environment. Generative tools can be used to create manipulated images, fabricate sexual content, impersonate people and intensify harassment. Women journalists who report on politics, gender, corruption or other contentious subjects can be targeted in ways that attack both their professional credibility and personal dignity.
UNESCO's May 2026 report, Tipping Point: Online Violence Impacts, Manifestations and Redress in the AI Age , produced with UN Women and the International Center for Journalists, documents the effects of these attacks.
In its 2025 survey, 45 per cent of women journalists and media workers said they had self-censored on social media because of online violence. Twenty-two per cent reported self-censorship at work. Five per cent said they had been targeted through deepfakes or manipulated images and videos, while nine per cent reported the non-consensual sharing of personal or intimate images.
The report also found that one quarter of surveyed women journalists had been diagnosed with or treated for anxiety or depression linked to online violence.
These are findings from a separate global study, not from the FAJ survey, and the two sets of figures should not be conflated. But together they demonstrate why the gender implications of AI cannot be treated as an optional addition to newsroom policy.
A woman journalist targeted with fabricated sexual imagery may face reputational damage, threats and pressure to withdraw from public discussion. If her employer does not respond, she is left to manage the consequences alone. If the newsroom treats the abuse as a personal social media problem rather than a workplace safety issue, it misses the effect on her ability to do her job.
News organisations need clear procedures for reporting online abuse, preserving evidence, assessing threats and providing legal, psychosocial and practical support. They should also have enforceable policies against sexual harassment and workplace discrimination.
The responsibility should not fall on women journalists to become more resilient while institutions remain unprepared to protect them.
Africa already has a framework for responsible AI. Implementation is the test.
The African Union's Continental Artificial Intelligence Strategy, endorsed in July 2024, provides a regional foundation for this discussion.
It recognises AI's potential to support development, create jobs and improve access to information. It also identifies risks including discrimination against women and vulnerable groups, job displacement, disinformation, privacy violations and surveillance.
The strategy calls for an African-centred, ethical and inclusive approach to AI development and use. It recognises that the continent needs investment in infrastructure, skills, research and innovation to benefit from the technology.
For journalism, however, a broad AI strategy must translate into protections that are specific to the newsroom. The African Commission on Human and Peoples' Rights'2019 Declaration of Principles on Freedom of Expression and Access to Information in Africa already provides relevant standards. It addresses media independence, the protection and safety of journalists, access to information, privacy and freedom of expression online.
These principles are not a substitute for employment law or newsroom policy. They do, however, make clear that technological change must be considered alongside fundamental rights and the public's ability to receive information.
The Addis Ababa action plan should therefore be judged by what follows the meeting. FAJ, its member organisations, the African Union, governments, media employers and development partners need to turn the commitments into practical measures.
There are at least four areas where that work should begin.
Media employers must put workers at the centre of AI decisions. Newsrooms should develop written AI policies with journalists and editors. These should cover verification, disclosure, source confidentiality, copyright, data protection, accountability and the use of AI in employment decisions. Workers should be consulted about changes to job descriptions, staffing and workloads.
Decent work must be treated as part of media freedom.* Fair and regular pay, written contracts, social protection, safe working environments and the right to organise cannot be postponed until news organisations become more profitable. Freelancers, stringers and contributors should not be excluded from protections simply because they are not permanent employees.
Gender-responsive safety measures must be enforceable. Media organisations should publish and implement policies addressing sexual harassment, online abuse, discrimination and unequal opportunities. Women journalists need accessible reporting channels, protection from retaliation and support when technology is used to target them.
The continental action plan needs measurable accountability. FAJ and its partners should publish clear responsibilities, timelines and progress updates. Governments should review whether labour and media laws protect journalists in changing digital workplaces. Media owners should report on how AI affects staffing, training and working conditions. Journalists' unions should ensure that freelancers and younger workers have a voice in negotiations.
Without public reporting, the action plan risks becoming another declaration of good intentions.
The public has a stake in what happens next
The future of journalism is often discussed in terms of technology, business models and competition for audiences. Those questions matter, but they are not the whole story.
Journalism is also a public service. People rely on it to scrutinise public spending, expose abuse, report on elections, explain policy and make information available to communities that might otherwise be ignored.
When journalists are underpaid, threatened or pushed into silence, the public loses access to information. When women leave newsrooms because of harassment or exclusion, the range of voices shaping public debate becomes narrower. When AI-generated material is published without proper checks, the credibility of the entire newsroom is put at risk.
The FAJ survey found that 47 per cent of respondents had considered leaving journalism. It also reported that 44 per cent experienced frequent burnout and 72 per cent said their workplaces provided no mental health or counselling support.
These findings raise a question that should concern media owners and policymakers alike: who will do the reporting if the people currently doing it cannot afford to stay?
AI may help journalists work differently. It may improve some newsroom processes and make certain forms of reporting more accessible. But it cannot replace the need for fair employment, editorial independence, professional judgement and protection from violence.
The measure of responsible AI adoption should not be how quickly African newsrooms introduce new tools or how much content they produce. It should be whether the technology helps journalists do better work without stripping them of their rights, livelihoods or professional independence.
Africa does not simply need newsrooms that can use AI. It needs newsrooms in which journalists can afford to work, question power, protect their sources and report without fear.
A technology revolution that leaves journalists insecure will not secure the future of African journalism.
