By Iduh Onah
THOUGH the emergence of Artificial Intelligence (AI) has profoundly altered journalism, particularly how newsrooms operate in their reporting and editing styles, the value of human moderation remains paramount.
This was the central message at the official presentation of “Newsroom AI Policy Guidance Resources”, a package of three documents produced by a non-profit, non-governmental media-oriented organisation, Safer-Media Initiative (SMI), under its Artificial Intelligence and Media (AIM) Project: Navigating the Present, Shaping the Future, held on Friday, August 7, 2026 at Savannah Suites and Resort, Area 3, Garki, Abuja. The theme came out very clearly in both the opening remarks and the review of the documents that followed.
The two interventions, while approaching the subject matter from different perspectives, were, however, generally complementary. While the opening remarks by Mr Peter Iorter, Executive Director/CEO of SMI, established the scale and urgency of AI’s ‘disruption’ of journalism, the review by veteran journalist and civil society activist, Dr Lemmy Ughegbe, interprets the ‘disruption’ using a vivid parable that portrays the danger of trusting a robot simply on the basis of its seeming accuracy.
Fundamentally, Mr Iorter used his remarks to frame the AI conversation around a simple question sequence: “What? So What? Then What?”
In answer to the question sequence, Iorter asserted the incontestable fact that AI is here and already transforming journalism with AI-powered tools enabling publishers to “gather, produce, edit, and distribute news at unprecedented speed.” He noted that at the same time, bloggers, influencers, citizen journalists and other content creators are competing with traditional media organisations for audiences, attention and credibility. “The newsroom we once knew is rapidly evolving,” Iorter observed.
The figures used to illustrate this transformation were equally striking. Citing UNESCO, Iorter noted that more than 54% of global advertising revenue is now concentrated on digital platforms, while 40% of users already utilise AI to create or modify content. These statistics, Iorter argued, demonstrate not merely the growing presence of AI but “the urgency for newsrooms to respond strategically.”
Iorter noted that in order to cope with the dynamics of transformation, the AIM Project, had already supported 107 journalists across 70 media organisations in Nigeria, empowering them with knowledge and practical skills for responsible AI integration into editorial workflows.
Although Iorter did not provide specific percentage measurement of AI usage across Nigerian newsrooms, he however hinted on a wider AI-adoption trend in Nigeria’s newsroom ecosystem. It is in that context that the AIM Project, Iorter says, postulated that the issue before the industry is not simply how many journalists are using AI, but whether the growing use is taking place within responsible institutional frameworks.
From Disruption to Responsibility
Iorter traced the ‘disruption’ to Clayton Christensen’s Theory of Disruptive Innovation, introduced in 1995. The theory, Iorter explained, describes how emerging technologies transform industries and force established institutions either to adapt or risk becoming obsolete. AI, he said, is “one of the most significant disruptive innovations of our time”, although he neither portrayed it as a technological villain nor as a miracle solution.
He acknowledged AI as a tool that creates opportunities for speed and efficiency, even as it raises concerns about editorial independence, misinformation and disinformation, deepfakes, AI-generated content and workforce displacement.
Economically, Iorter argues, AI-driven advertising technologies are also accelerating the shift away from traditional newspapers by reducing print circulation, moving advertising revenue towards digital platforms and intensifying competition for audience attention.
The AIM Project’s proposed response, therefore, is neither rejection nor blind enthusiasm. “The answer is not to resist disruption but to respond to it responsibly,” Iorter counselled, noting that that response is anchored in the human element built on traditional values of journalism to ensure accuracy, fairness, accountability, transparency and editorial independence. In other words, the desired relationship between human and artificial intelligence is one in which “technology enhances journalism without replacing its ethical foundations.”
It is in this context, Iorter argued, that the AIM Project conceptualised the three “Newsroom AI Policy Guidance Resources” components: “Editorial Policy Guidance on Artificial Intelligence for Nigerian Newsrooms”; “AI Readiness Guide for Newsrooms in Nigeria”; and the “Gender-Sensitive AI Policy Guidance for Nigerian Newsrooms”.
The Parable of a Kingdom Troubled by Theft
If Iorter’s opening remarks offered the rough estimation and strategic framework, Dr Ughegbe’s review created the perfect imagination for practical utilisation of AI in the newsroom. Ughegbe began not with another statistic but with the parable of “a kingdom troubled by theft;” a powerful metaphor for the central dilemma of AI-assisted journalism.
A kingdom had a king who is impatient with slow justice. He heard of an oracle that could identify any thief instantly. When summoned, the oracle, a tall bronze mirror, is “calm and certain, never once pausing to think.”
The king was delighted. Cases that had previously taken weeks were resolved in moments. Grain thieves, cattle raiders and market cheats were identified and punished before sundown. The kingdom marvelled at the oracle’s speed.
Then came the missing merchant’s strongbox. Asked who had stolen it, the oracle immediately named the weaver’s wife. She was arrested and brought before the court. Only then did the merchant’s son confess that he himself had stolen the money to settle a gambling debt.
When the king later demanded to know why the oracle had falsely accused an innocent woman, its answer captured the entire AI dilemma in one sentence: “You did not ask me if I was certain. You only asked me to be quick.”
It is an especially effective parable for journalism, and even more so, AI-assisted one, because newsrooms are themselves under pressure to be quick. The race for breaking news, clicks, audiences and digital relevance can make speed appear synonymous with success. However, journalism’s professional and social responsibility requires something more than speed.
In the kingdom, the king’s solution was telling. He kept the mirror because it was “rarely wrong and never wasted a moment”, but he insisted that before any accusation left his court, a human voice had to ask the question the mirror could not ask itself: “are you sure?”
Ughegbe then made the connection explicit: “That is the choice before Nigerian journalism today. Artificial intelligence is our new oracle: fast, confident, rarely wrong, and never built to doubt itself.”
The parable turns an abstract policy discussion into an easily understandable editorial principle: AI can assist with the work, but the journalist must retain the capacity as well as the duty to question or moderate the machine.
For Ughegbe, the point therefore is not the mere adoption of AI simply because competitors are doing so, or because it is fashionable, or because management believes it will reduce costs. Rather, newsrooms should ask what problem a proposed AI tool is intended to solve, whether the problem can be solved without AI and what could happen if the system fails.
The review recommends beginning with relatively low-risk applications such as transcription and translation, while exercising greater caution in investigative journalism, conflict reporting and fully automated content production. It also stresses training because an AI-generated response may have the sound and appearance of authority while being entirely false.
Ughegbe illustrated the point with another memorable example: a lawyer who submitted a court brief containing six apparently authentic previous rulings, complete with judges’ names, docket numbers and quotations, only for it to emerge that none of the cases existed. The lesson for journalism, Ughegbe said, is direct: “Confidence is not the same thing as truth.”
That assertion remains an enduring caution for all newsrooms and journalists using AI. For Ughegbe, a fabricated quotation produced in seconds can be more dangerous than an obvious error because its polished language may encourage a hurried reporter to accept it. He therefore calls for more than theoretical awareness, suggesting practical instruments, including a risk-classification matrix and an approved-tools register, to turn policy principles into everyday newsroom practice.
The Machine Versus Newsroom’s Liability
Perhaps the strongest editorial principle in the documents, Ughegbe said, is the insistence that responsibility cannot be outsourced to technology. Referencing the “Editorial Policy Guidance”, he stated: “The newsroom retains full legal and ethical responsibility for all published content, regardless of whether it is created by humans, AI systems, or a combination of both.”
That principle, Ughegbe asserts, has profound implications, noting that if an AI-assisted report is inaccurate, defamatory or discriminatory, the newsroom cannot point to the machine as though it were an independent employee capable of accepting professional or legal consequences. AI cannot defend a reputation, appear before an editorial board or answer questions in court, and therefore, the responsibility remains with the reporter, editor and media organisation.
Ughegbe thus insists that AI-generated or AI-assisted material – facts, figures, quotations and framing must – must undergo rigorous human check before publication. “Human oversight without actual verification is merely ceremonial oversight,” he said. In other words, having a human being somewhere in the workflow is not enough. Human oversight has meaning only when the human actually interrogates, verifies and takes responsibility for the output; what is called gatekeeping in journalism.
What Reporters Must Never Feed AI
The review, turning attention from what AI can produce to what journalists should never feed it, strongly warned against putting confidential information, unpublished material or sensitive source data into public or third-party AI systems. Ughegbe noted that for investigative journalism, the consequences could be severe.
According to him, an attempt to save time by uploading confidential documents to an external AI platform could expose an investigation, compromise a source and create legal risks for the media organisation.
His proposed remedy is a newsroom’s “Do Not Input” checklist, an instruction in AI-era journalism that he said is synonymous to what “off the record” has been to traditional journalism in the effort to protect sources.
The Gender Question and Algorithm
In his commentary on the third document: “Gender-Sensitive AI Policy Guidance”, which further broadens the AI discourse on the premise that technology is not automatically neutral such that AI algorithms can reproduce inequalities embedded in their training data and in the institutions that deploy them, Ughegbe points out that women journalists already experience disproportionate sexualised abuse, threats and doxing, while generative AI introduces additional dangers including manipulated photographs, voice cloning, synthetic sexual content and deepfakes.
In praising Gender-Sensitive AI Policy Guidance’s suggestion that: “The newsroom adopts a zero-tolerance approach to all forms of gender-based violence, harassment, intimidation, and discrimination, including harms enabled, amplified, or automated through AI systems and digital technologies,” Ughegbe praises the Guidance for moving beyond condemnation to concrete institutional responsibility. He identifies confidential reporting channels, protection from retaliation, legal assistance, psychological support and disciplinary consequences, with protections extending to freelancers, contributors and fixers.
The same concern, he says, applies to less obvious forms of algorithmic bias. If an AI system repeatedly recommends male experts for politics and economics but women mainly for health and family stories, it could reproduce old stereotypes while creating the impression that its recommendations are objective. In other words, artificial intelligence can automate bias just as efficiently as it can automate tasks.
From Documents to Newsroom Culture
Taken together, both Iorter and Ughegbe point towards a transformation that is larger than the publication of three newsroom policy documents, with Iorter advocating for a strategic case for AI readiness while Ughegbe calling for a human-guided and responsible AI use.
In other words, Iorter, or the AIM Project (through its documents), asserts that AI is already reshaping the industry and provides statistics indicating the broader scale of digital and AI adoption. On the other hand, Ughegbe’s review projects what happens when the newsroom becomes overwhelmed by the AI oracle’s speed and confidence in processing journalism.
The shared response to the evolving problematic is not technological retreat but institutional maturity and intentional or conscious preparedness. It is these mindsets that will facilitate the three documents into “a practical roadmap for newsrooms” seeking to integrate AI while preserving editorial independence, protecting journalists and sources, strengthening accountability and sustaining public confidence.
Their success, Iorter stressed, will depend not simply on publication but on formalisation: “Their true value lies not in their publication, but in their adoption and application.” Putting it in simpler terms, the AIM Project produced the documents with the aim of aiding Nigerian newsrooms to constructively apply them in their operations.
Ughegbe illustrated this point even more clearly when he said: “A policy that remains on a shelf cannot prevent an AI-generated falsehood. It cannot protect a confidential source. It cannot assist a journalist facing a synthetic sexual attack. It cannot preserve public trust.”
The AIM Project’s ultimate projection is that both contemporary and future newsroom is neither man against machine nor machine against man but responsible coexistence. Thus, just as the future of journalism cannot be secured by pretending that AI can be wished away, so also the value of human judgement cannot not be substituted by AI.
