As AI revolutionizes journalism practices worldwide, newsrooms in Nigeria increasingly are integrating new such tools to enhance storytelling and fact-checking.
These AI tools, although unable to replace the work of humans, can handle a wide variety of tasks. From summarizing and analyzing large datasets, to verifying information, the new technology is indeed shaping and changing how newsrooms in Nigeria operate.
Here’s a closer look:
Debunking false information
The internet has made it possible for people to access and share information swiftly. This information is often disseminated on social media today, where false news spreads six times faster than factual news.
But false information doesn’t only spread on social media. It also travels by radio, noted Silas Jonathan, research and digital investigations manager at Dubawa, a West African fact-checking project initiated by the Centre for Journalism Innovation and Development. Recognizing this, in May 2024 Dubawa introduced an AI-powered chatbot and an audio platform to debunk false information.
“When we talk about fact-checking mis-and disinformation, we look only at social media. But we realized that people lie a lot on the radio and can say something misleading but nobody follows up to check,” Jonathan explained. “We decided to create [the Dubawa Audio Platform] that will [help journalists] monitor and identify false claims [on radio] and debunk them.”
After registering, users can use the Dubawa Audio Platform to monitor live radio programs, transcribe audio recordings, and verify claims.
Unlike most AI tools today, which are created with a Western orientation, the tool allows journalists to transcribe local audio recordings to text in Ghanaian and Nigerian English dialects, and Pidgin English. The organization also has plans to expand the platform into local languages in Ghana and Nigeria, as well as Arabic and French, Jonathan noted.
The Dubawa Chatbot, meanwhile, operates on WhatsApp, generating responses to queries and claims users submit. The chatbot also produces relevant links users can visit to get more information, explained Jonathan.
The responses are informed by fact-checks conducted by Dubawa and the International Fact-checking Network. “When you make a query, it’s not just giving you random answers from unknown sources. It’s giving you feedback and references from already verified information that followed through the editorial process,” he said.
One challenge Jonathan pointed out is the delay in getting verified information not available in the chatbot’s database. “When you ask [the chatbot] a query that is not yet verified, you don’t get any viable answer. [The response] has to have been fact-checked or written by credible media that we have selected to serve as the database,” he said.
Overall, the number of users taking advantage of the tool is increasing. Just last month, the chatbot recorded 9,000 users and Dubawa Audio 4,000 across Ghana and Nigeria, according to Jonathan.
Turning data into stories
In 2022, Dataphyte, a media, research, and data analytics organization in Nigeria, launched an open-source AI tool called Nubia that helps users analyze large datasets and turn them into stories.
These stories are a “first draft” that human editors need to fine-tune, explained Ibilolia Akahome, a programs and business development manager at Dataphyte. “This tool was created as a synergy between artificial intelligence and human intelligence,” she said. “The first draft must be improved on by the journalists.”
Bernardo Motta, associate journalism professor at Roger Williams University, echoed the importance of human oversight of the tool. “What makes stories good and appealing to people is that human connection. When we are talking about journalism, in which you talk about real-world people, AI does not have that information, or when it does it distorts that information,” Motta said. “It becomes a problem when the AI makes a conclusion.”
Nubia has helped Oluwasegun Abidoye, a data analyst and an investigative journalist with Dataphyte, produce a variety of stories over the past year. In July, for instance, he used the tool to report on whether customers were getting the value they paid for from electricity distribution companies in Nigeria.
“I input the dataset and it gave a real-life story. It was really useful and user-friendly, and it saved time,” said Abidoye. “With Nubia, we were able to know whether they were supplying enough power, and even the total revenue.”
One major challenge Abidoye encountered, however, is that he wasn’t able to input multiple datasets at once to compare against each other. “We wanted to compare [findings] with another country, but the initial dataset could not help us compare with other countries. We had to generate another dataset,” he said.
Understanding how AI tools work
To encourage users, both Dubawa and Dataphyte are investing in training. This year alone, Dataphyte has held more than 20 training sessions for media professionals, lecturers and university students, according to Akahome, while Dubawa has trained about 4,000 journalists across Africa, in The Gambia, Kenya, Liberia, Nigeria and South Africa on how to use its AI tools.
Given the ethical concerns arising from the use of AI tools in newsrooms, Motta argued that journalism ethics should be part of the conversation during training sessions.
“We never can assume that people who are doing journalism have been trained on the ethics of journalism, so that has to be part of the training. Journalists need to understand how AI tools are created and how they function. If you're just trained how to use a tool, you are substitutable,” he said.
Journalists must be aware of data that AI tools may be collecting from them, too. “If you are using a tool that you don't know what it’s doing, I can guarantee you that it is collecting information from you, and that's another problem for journalism ethics, because that can give information about sources,” Motta added.
Yutong Liu & The Bigger Picture / Better Images of AI / AI is Everywhere / Licensed by CC-BY 4.0.
Leia mais artigos de
