The Guild of Corporate Online Publishers (GOCOP) has organised a training session for its members on the effective use of Artificial Intelligence (AI) tools to improve efficiency, productivity and profitability in online publishing.
The training was held on the sidelines of the Guild’s 10th Annual Conference in GRA, Ikeja, Lagos, with the theme, “2027 Elections: Trust, Transparency and Shared Responsibilities”.
The session, titled “The AI-Powered Newsroom: AI and Prompt Engineering for Online Publishing”, was facilitated by AI expert and publisher of The Lagos Times, Biodun Durojaiye.
Durojaiye said the training was designed to help publishers understand how AI could be deployed to produce content faster, improve quality and optimise different stages of the publishing process.
He defined AI as technology that enables computers to perform tasks that ordinarily require human intelligence, noting that the publishing industry had evolved significantly from the traditional newsroom-newspaper-reader structure.
According to him, AI can understand language, recognise patterns, analyse information, generate content, solve problems, make predictions and assist decision-making, adding that users do not need to be programmers to benefit from the technology.
Durojaiye explained key AI concepts, including Generative AI, which can create text, images, audio, video and code, as well as Large Language Models, which are trained on large volumes of information and can understand and generate human language.
He cited ChatGPT, Gemini, Claude and Copilot as examples of tools powered by large language models.
The facilitator also distinguished AI from conventional search engines, noting that search engines primarily provide links and sources for users to investigate and assess, while generative AI can process, transform and analyse information.
He, however, warned that AI could produce inaccurate information and therefore users must independently verify information generated or sourced through AI.
Durojaiye also introduced participants to prompt engineering, which he described as the art and science of giving AI clear and effective instructions to produce useful results.
He contrasted a basic prompt such as “write a story about the budget” with a more detailed instruction requiring AI to act as a senior Nigerian newspaper correspondent, analyse a budget speech and produce a 600-word report using the inverted pyramid structure, with the lead focusing on the most significant development.
He outlined a five-part prompt formula comprising Role, Task, Context, Requirements and Output, stressing that the quality of instructions given to AI would influence the quality of its response.
Speaking on the AI-powered newsroom, Durojaiye said the technology could assist with news gathering and research by analysing lengthy speeches and reports, summarising government documents, extracting key facts, comparing documents, generating interview questions and identifying potential story angles.
He added that AI could analyse datasets, conduct background research, simplify complex materials and support investigative work, while stressing that it could not replace journalism or the investigation process itself.
According to him, AI could also support various stages of the publishing workflow, including news gathering, research, interviews, transcription, writing, editing, search engine optimisation and social media and newsletter analytics.
He said one news event could be transformed into several content products with the aid of AI, including news reports, features, explainers, profiles, interviews, opinion drafts, headlines, captions, newsletters and press-release adaptations.
However, Durojaiye stressed that while AI could accelerate content production, journalists remained responsible for facts, context and editorial judgement.
He also highlighted the use of AI for search engine optimisation, noting that online publishers needed to ensure their stories could be discovered by the right audiences.
He said AI could assist with search-friendly headlines, SEO titles, meta descriptions, keywords, search intent, frequently asked questions, content structures, internal-link suggestions and content clusters.
“The story isn’t finished when you publish it. It is finished when the right audience can find it,” he said.
