A survey presented at a training workshop organised by the Safer-Media Initiative (SMI) in Lagos on Wednesday, 18 March, shows that while 95 per cent of journalists in Nigeria use artificial intelligence (AI) tools at least once a week, only 10 per cent have received structured training from their newsrooms.

Organisers warned that the gap exposes journalists to risks around accuracy, ethics and job security, calling for urgent newsroom support.

Themed “AI Tools for Journalists: Effective, Responsible, and Safe Use,” the training  and the survey form part of SMI’s activities under the Artificial Intelligence and Media (AIM) Project, supported by UNESCO’s International Programme for the Development of Communication. The session was moderated by DevReporting’s Programme Manager and Associate Editor, Samson Ademola.

High AI use, low training

The Executive Director of SMI, Peter Iorter, said the findings point to a growing imbalance in newsrooms.

“Journalists are already using AI tools regularly, but very few have been trained by their newsrooms on how to use AI effectively and responsibly,” he said.

According to the survey, 95 per cent of respondents frequently use tools such as ChatGPT, often without guidance on verification, safety or ethical standards. Mr Iorter warned that widespread use without structured support could weaken the quality and credibility of journalism.

Skills gap, not AI, will lead to job displacement

Contrary to concerns by some journalists that AI will take over their jobs, Mr Iorter argued that the real risk lies in the skills gap.

“Another thing we found in the survey was that journalists responded that the greatest threat to AI is job displacement. Journalists think that AI will take up their jobs. AI is not going to take away journalists’ work but journalists who fail to adopt AI tools effectively may have their jobs taken away by those who have prepared themselves adequately,” he said.

Mr Iorter noted that journalism is undergoing rapid transformation, driven by changing audience preferences, evolving formats and new methods of news gathering, processing and distribution.

“This disruption is caused by technology, and artificial intelligence is one of such technologies, such that even the traditional media outfit have no option than to adapt their ways to the changing narratives,” he said.

According to him, the growing influence of AI makes it necessary for journalists to build the skills required to use the technology responsibly and effectively.

AI already reshaping newsroom work

The trainer and founder of Generative AI Journalism with Titi, Titilope Fadare-Oparinde, demonstrated how AI tools are already embedded in newsroom workflows.

She said journalists now rely on AI for transcription, translation, summarising reports and repurposing content. According to her, the debate has shifted from whether to use AI to how it should be used responsibly. “The question is no longer whether to use AI, but how to use it properly,” she said.

She emphasised effective AI prompting as central to obtaining useful AI outputs.

Participants charged on ethical and safe AI use

Mrs Fadare-Oparinde warned about the limitations of AI, particularly its tendency to produce inaccurate or fabricated information, commonly described as hallucinations.

“AI systems may not admit uncertainty and may generate convincing but inaccurate facts, quotes or statistics. All AI-generated content must be reviewed, edited and verified by human journalists,” she said.

She stressed that editorial judgement cannot be delegated to AI, particularly in decisions involving sources, context and representation. Participants were advised to treat AI as a support tool rather than a replacement for professional judgement.

Mrs Fadare-Oparinde cautioned participants against exposing sensitive information to public AI tools, with warnings that confidential materials such as unpublished investigations, source communications and leaked documents could be stored or reused by such systems.

She urged that personal or official email addresses should not be directly linked to some AI tools with questionable or untrusted innovation sources. She suggested that participants create a new email address for such AI tools, working from personal notes and using controlled tools, while reinforcing that AI must never be used to identify or handle anonymous sources.

She also examined broader risks, including gender bias and the growing threat of AI-enabled harassment, particularly targeting women journalists through deepfakes, voice cloning and impersonation.

Mrs Fadare-Oparinde also expressed the need for transparency with audiences, recommending clear disclosure of AI-generated content, while maintaining full editorial accountability for published work.