The Transparency Paradox: An ethnographic investigation of AI Ethics and Accountability in Nigerian journalism
TL;DR
The study found that GenAI provided efficiency gains with tasks where it worked well but led to “double work” with tasks where it failed, and reported a transparency paradox, where the journalists express belief in disclosure but find it difficult in practice.
Abstract
The rise of artificial intelligence has led to disruptions across major knowledge work, including journalism. The recent proliferation of generative artificial intelligence (GenAI) and large language models (LLMs) such as ChatGPT, Gemini and MetaAI, has led to an increased study of the influence of AI in journalism. However, much of these studies have been focused on western contexts, with very little knowledge of how AI is transforming journalism in Africa. To begin addressing this gap, we conducted a two-phase study to understand how Nigerian journalists are using AI. Nigeria is not only significant as Africa's largest democracy but also as one of the first countries on the African continent with a national AI strategy that recognises ethics and transparency as concerns in the deployment of AI in knowledge work. Our study combined semi-structured interviews with 20 journalists and an ethnographic investigation within a Nigerian newsroom. The study found that GenAI provided efficiency gains with tasks where it worked well but led to “double work” with tasks where it failed. We also report a transparency paradox, where the journalists express belief in disclosure but find it difficult in practice. We discuss the implication for journalists, AI companies and policy makers in ensuring effective AI governance in an African context.