The transformative power of generative AI is shaking up the worldwide media industry, with immense consequences for the Global South.
A few weeks ago, I had this immense opportunity to participate in a virtual discussion hosted by a reputed international institute on the topic of AI and the future of the news. The program discussed how media houses worldwide are utilizing new emerging technologies and generative AI tools and how they will impact the future of the news.
The program focused on various tools, techniques, and algorithms, the importance of AI guidelines, and how artificial intelligence (AI) will shape the future of news and journalism around the globe. Around 30 journalists from all over the world with diverse backgrounds and expertise attended the program. Being the only participant from Nepal, my concern was how we can use AI in the newsroom of small countries like Nepal or the other countries in The Global South. The six-hour-long discussion program superficially highlighted that there's uneven progress in integrating artificial intelligence language models across various regions from the West to the Global South.
Though we are completely in the initial stage of implementing the generative AI tools in Asian newsrooms, the experiment of AI technologies and companies in languages has also been limited. However, some of my fellow participants from Africa provided intriguing examples of utilizing generative AI tools to generate headlines, and captions for photos, and curate content for visual stories.
Despite the demographic differences among the participants, we all had similar concerns about using AI in the Newsroom. That is “Trust.” How can journalists build and maintain trust with the information that is being provided by generative AI? There's an intense fear that AI could become a mere cost-cutting tool for news outlets, churning out vast stories devoid of the crucial human oversight that ensures accuracy and ethical reporting.
Taking investigative journalism to new heights
The concerns about the impact of AI in the future of news and journalism are not merely hypothetical. Western media outlets have already downsized the journalism workforce after incorporating AI technologies. This fear is spreading rapidly to the global south, where the true impact of AI on journalism is still unseen. Will AI encroach upon the roles traditionally held by human journalists and reporters? In a bid to cut costs, resource-challenged media companies might resort to AI without due consideration, but such an approach could have harmful effects on journalistic integrity, the diversity of perspectives that enrich news coverage, and the important context of local languages that skilled reporters provide. Moreover, the potential for biased AI systems to perpetuate inaccuracies further fuels these concerns, especially in the areas where technologies are limited.
During the sessions, professors and leading experts in machine learning emphasized the contrasting views surrounding AI. They acknowledged the complexities and ongoing debates within the big tech companies of AI itself. In contrast, while members of society view generative AI with some hopefulness mixed in their caution, journalists naturally feel uneasy about its potential implications. They stress the importance of powerful policies, to be formed guiding safe practices around implementing this technology into journalism's structure in a responsible manner. Moreover, they emphasized the importance of using AI in collaboration with media houses, holding journalistic ethics, editorial judgment, and accuracy, regardless of circumstances.
However, developing nations face high stakes amidst the changing technological landscape. Language models such as GPT-3 predominantly rely on data from major languages, particularly English resulting in significant inaccuracies when applied to those with low digital presence. The complexity of linguistic structures, idioms, and cultural contexts important for accurate translations poses a challenge for these models. Additionally, the training data for these AI models often skews heavily towards content from Western, Educated, Industrialized, Rich, and Democratic (WEIRD) societies. This inherent bias can lead to the marginalization of important perspectives and issues relevant to the global south.
News and stories translated through such models risk filtering out crucial context and cultural nuances, potentially distorting the understanding of critical events and issues. A well-known Panamanian journalist has also discussed the limitation of this technology in interacting with the experts in Norway, Georgia, and Ghana.
In the year 2024 when several prominent fact-checking organizations like FullFact, Maldita, and Newtral are implementing AI tools to facilitate their work, outlets, in small countries in the Global South with minority languages, newsrooms only have questions about whether AI tools are accurate to help their fact-checking efforts? And the answer is NOT YET. The lack of specialized knowledge in minority languages of developing nations poses a huge challenge to AI's ability to comprehend complex journalistic terminology, events, and reporting protocols. This challenge increases the risk of misinformation and disinformation when dealing with sensitive news related to politics, policies, marginalized communities, and conflicts.
Meanwhile, media industries in the West are exploring AI technologies to enhance the quality of journalism by taking on mundane tasks such as data analysis and draft writing. Large news outlets like The Washington Post have already started implementing AI tools to streamline their processes. With significant financial resources at their disposal, major media companies in the West are well-positioned to integrate AI while maintaining journalistic standards effectively.
However, the same is not true about news organizations across Africa, South Asia, and Latin America. Since AI technologies are not prioritizing language automation, proper oversight mechanisms, or skilled engineers to maintain credible pieces of information, there are increasing risk factors associated with potential errors which could lead to disinformation. Experts suggest that big AI companies should research and improve deep learning processes in the markets of the Global South rather than ignoring the local context and language anomalies.
If generative AI technology freely spreads throughout Asian newsrooms, there will be a fear of a "credibility crisis." We have already seen the rise of misinformation via a few chatbots on social media like Chat Gpt, Quillbot, and Google Gemini. Without proper policy and governance measures in place, the risk exists that mainstream news sources may unintentionally promote fabricated stories generated by these AI technologies. The concern goes beyond accuracy; it's also reasonable to worry about inadequate representation shaping this type of system since Western perspectives could be favored over vital nuances relevant to cultural interests, traditions, and accounts from regions globally outside their sphere of understanding.
In the newsrooms of the Global South, incorporating AI models that rely primarily on English language data will embed a structural bias that obstructs the telling of diverse narratives. This could cause severe harm to societies already grappling with their voices being marginalized in global conversations.
There has been a debate about the accelerated implementation of AI in the newsrooms of developing countries. Instead, the policymakers should push for increased access and involvement in constructing and training foundational language models. Doing so could lead to more just and inclusive systems that reflect our values as a society. AI experts say, “AI is not an enemy of humanity.” We require AI technologies to make our tasks easier; however, these technologies should include all language models and research all the information from all over the world to maintain accuracy, balance, and credibility.
For now, the language model of the small countries of the global south must be embedded in the large setting of generative AI tools to be competent with global newsrooms. Since countries like Nepal are undoubtedly rich in culture, tourism, nature, and history, and even have wonders of the world, big tech companies need to look over the language models and resources.
While there are concerns about journalists’ job loss due to rapidly advancing AI capabilities, experts from various tech companies say, “Generative artificial intelligence also offers opportunities for journalists.” If AI is used responsibly, it can be used for the storytelling of the voices of people previously unheard. Moreover, customized algorithmic approaches designed with specific local communities’ needs can aid struggling journalists who face resource constraints but possess unique insights into their region’s demographic nuances.
The transformative power of generative AI is shaking up the worldwide media industry, with immense consequences for the Global South. It's uncertain whether this will create a chance to cultivate information ecosystems that are empowering and representative or if it'll lead to entrenched inequalities flourishing in new forms of technology. The decisions made today by journalists, technologists, policymakers, and civic leaders will ripple across the generations to come. Responsible step over AI's news media applications is paramount to fostering a free, pluralistic, and equitable public discourse worldwide.
