AI is a double-edged sword. It can be a powerful tool that makes our jobs easier but it can also spread misinformation, manipulate public opinion, and erode trust in journalism.
Former CNN anchor Zain Verjee and intelligence expert Candyce Kelshall, have shared a blueprint to help global newsrooms fight misinformation. This handbook—Election Interference and Information Integrity: A Newsroom Blueprint—provides media professionals with practical guidelines and AI-driven tools to strengthen the integrity of their reporting, particularly covering elections.
TechCabal spoke to Zain Verjee about the motivation behind the handbook, the threat of misinformation and disinformation exacerbated by AI, and the ethical considerations for AI use in African newsrooms.
(This interview has been slightly edited for length and clarity)
TC: What would you say is the story behind this handbook?
I think one of the biggest threats in the world today is misinformation and disinformation, and because the world is so polarised, and we’re seeing it with the US elections, and we’re seeing it coincide at a time where artificial intelligence and new technologies have made it so easy to create fakes. And it comes at a really important time for journalists, where, at the same time, trust is at an all-time low with many journalistic establishments and audiences. And so the idea of this came about in the sense that I’m a former journalist, and I’m looking at this going wow, as if I was in a newsroom, what would I do? How would I be thinking about a lot of these really important topics? How would I maintain trust with my audiences, and how should I understand the evolution of technology and artificial intelligence around me?
And I felt that there was no place where journalists could comfortably go to one resource that really analysed artificial intelligence and that looked at newsrooms as information hubs, as intelligence hubs, with the principles of determining what’s real and what’s not real that are applied in intelligence agencies that I’ve worked with in the past and used as sources should be applied to journalists and newsrooms. So really, it was a combination of all of those things, a combination of my personal experience, my interest in artificial intelligence and technology, and a really difficult time in the world, and in particular for journalists covering elections.
TC: What context was this handbook built in mind? Was this specific to a particular region? Did you consider it peculiar to maybe the global context?
I think that the core learnings and the core ideas in the handbook are universal, right? Fact-checking is something universal. Sourcing multiple sources, understanding what is, determining a piece of video or audio or a photo and whether it’s real or not, isn’t just for one part of the world.
It’s what all of us are facing right now; deep fakes and manipulated media. And so the context of this is if you treat news and content as an intelligent product that requires assessment, that requires certification, that requires verification before being released to the public.
Newsrooms will increase their integrity all around the world, and we build trust. So it’s really been written with a universal context in mind. But as you know, I’m also a Kenyan and a journalist, and I do think that on our continent there are other challenges that we face, and there are probably aspects that could be built out to focus a little bit more on the context in multiple African countries. But this handbook universally addresses problems that we all face: critical thinking, digital literacy, misinformation, disinformation and information integrity, and best practices that can be applied across the board.
TC: What are the most important aspects of this handbook that speak to the context of African newsrooms?
I’m going to reinforce the misinformation and disinformation piece, right? Because understanding and having the ability to determine what is misinformation and what is disinformation simply requires making ourselves literate about these areas. Those are two completely different things, right? Missing. For example, misinformation is something that I agree with. It’s my worldview. I have an emotional connection. I share it with you. You agree with me. You share it. I’m not doing anything malicious, and I’m not trying to disrupt Nigeria’s elections, right? So I’m participating in amplifying misinformation.
I share things around on WhatsApp with my friends, right? Because we all believe it. As a journalist, this handbook is teaching all of us how to recognise misinformation and the intention behind it.
Disinformation is what bad actors are deliberately and maliciously creating information that is completely and utterly fake with the intention to destabilise like Russia’s current interference in the United States. CBS News recognised this as disinformation, so did the FBI, and the same with CNN.
This handbook allows journalists to basically have a checklist of what is misinformation and what is disinformation. So I think that’s really one of the most important things. In terms of the African context, there are a lot of considerations here. I think the first is our continent is mobile-first. And so a lot of verification tools are not on mobile. I think that that is a big challenge. If things are being shared on WhatsApp and TikTok, you have to have tools on mobile devices that can recognize what is real and what is not.
I think African newsrooms should be focusing on mobile-first verification and tools and processes that need to work on mobile devices. Some of the solutions would lie in creating, you know, for example, shareable fact checks that can work on low bandwidth connections, whether it’s using local communities, different language groups, SMS alerts, or different ways to create verification processes that make sense for us in our environment.
I would say also that we know that a lot of newsrooms on our continent have a lot of resource challenges and infrastructure challenges. So I would argue that collaboration is really critical, particularly around elections. For example, creating joint fact-checking databases, or working together during elections, having media houses team up during election periods to combat disinformation, various things like that, which comes down to processes in newsrooms that need to be introduced to improve verification, where you turn local knowledge and local cultural nuance into a verification process.
TC: Globally, there’s still generally a sense of skepticism around the use of artificial intelligence in newsrooms. How should African newsrooms think about AI?
I’m quite bullish about AI, but I’m also bullish about being human. I think that artificial intelligence is here to stay, and these tools are going to exist. They’re going to only become better and stronger and more powerful. By choosing to ignore it, you put yourself in jeopardy, both in terms of your future job and also just in terms of learning a skill set.
I have a very strong view that learning how artificial intelligence works, the kind of tools that can help make your process efficient, is a good thing. It doesn’t mean that tool is replacing you, the human being who has critical thought, what it’s doing, if deployed correctly, in my opinion, is that it’s just going to give you back your time. So you can call your sources, go and meet someone for a little bit longer. You know, think more carefully about the kind of story that you want to write.
I think that journalists who do not use AI will be replaced by journalists who do use AI. And I think that it’s not also rose-coloured glasses that I’m seeing things through. You know, there are AI tools that are not so great. There are AI tools that are not great for African contexts because those data sets come from white, Western resources. They scrape the internet, and that’s, you know, that’s what they have. They don’t have our local contexts in there, which is why we also need to build our own small language models and our own African dataset so we can use them to help us tell our own stories and our own languages, right, with our own cultural nuances.
But if we’re going to bury our heads in the sand and go, “This is no AI for me.” You are actively sidelining yourself from the future. So I don’t want African journalists and African talent to not critically be using tools and products and developing our own tools and products that are good for us, right? Because it’s just something a little scary, and it’s out there because it’s here, it’s coming, it’s going to stay. But that leads me to my second point, which is, I’m bullish on humans.
Nobody can replace our interaction here. We have a laugh. I see you. You kind of see me. You ask critical questions, you have historical knowledge, and you have empathy if you’re interviewing a subject. These are human things that a machine, at least for now, cannot replicate. You are the critical thinker and the last stop in a newsroom. So this handbook is a resource. The tools here are things to help you think. Help speed up. Is this real? Is this not real? Can we use this reverse imaging tool? What does it tell us about the metadata? Right? We’re intelligently leveraging technology, but we’re not lazily allowing technology just to take over.
TC: What would you say are some of the ethical considerations for the use of AI in the African context?
The first thing to understand is that because these tools are so powerful, you’re also handling sensitive data about real people. If you’re using AI in your newsroom whether it’s for translation, fact-checking, or content analysis, you really need to have some kind of in-house guidelines about data and protection. How are you going to store people’s information? Who has access to people’s information? How do you respect local community values and privacy? How you do how do you do that? And I think that it’s really incumbent on journalists themselves and leaders to set your own ethical standards for your newsrooms.
What is our newsroom policy about data handling? How do we deal with different communities when we’re collecting their data and their information, we’re using it in an AI tool that we’re developing, or inputting into analyse, you know, people’s social media channels? I think the onus is on us, on journalists, to set some ethical standards and guidelines. Two other things that come to mind are consent and transparency. You always want to be upfront. If you’re using an AI, or if you use something to help you write a story or develop a photograph on Midjourney, or if you use Claude or ChatGPT, you’ve got to be transparent with your audiences and say, “I use the following AI to help verify my stories. I use this AI to analyze some data.” Because your relationship is built on trust and transparency with your audience.
The last thing is, I do think there’s an issue with data sovereignty. African newsrooms should be careful about where your data is going when you use some AI tools. You may store data on tools that live on servers outside the continent. If you use any tool, a journalist should ask, “Who owns it? How is it going to be used? How is our data being stored?” All of these are ethical considerations that I think are really important.
TC: If any African journalist is reading this handbook today, what sections would you say are the most important that they must check out?
I would say critical thinking is the most important because it really puts the human in the driver’s seat and it forces us to look at sources in a way that we’re asking the key questions that allow us to determine, should that be published? Should that air? Is that thing viral because, you know, a large volume of people believed in it, and whether it’s not really true?
That would be the most important thing. It would be to ask yourself, who, why, what, when the context, bias, fact-checking, intention, and clarification. That is also how we evolve as human beings, as responsible citizens, and as ethical people who acknowledge the importance of artificial intelligence in the world today. But we’re not driven by it. We’re using our own gifts, our own talents, and our own knowledge base that’s cultural and contextual to make judgments.