Introduction
The Mona School of Business & Management is part of the University of the West Indies in Jamaica. We spoke with Maurice McNaughton, the Director of the Centre of Excellence for IT-enabled Business Innovation at Mona, who shared details about a grant-funded research project where NewsCatcher plays a key role. The project focuses on using generative AI to track and analyze gender-related narratives in the Caribbean.
Identifying the Problem
The goal of the project was twofold: first, to build out the local capacity for working with generative AI beyond simply using tools like ChatGPT, and second, to develop a practical application of this technology tailored to the Caribbean region. The team decided to build a "Gender Monitor," a tool designed to track and analyze narratives around gender issues in the Caribbean region. Initially, they considered social media platforms as data sources for their real-time insights, specifically Twitter, but concerns over the platform’s content reliability and the prohibitive cost of API access led the team to focus on mainstream media instead.
Build vs. Buy
Recognising the challenges which came with building a scraping system in-house, the team decided to focus on a solution that could be implemented with minimal setup while still offering the flexibility to meet their project’s goals. This “buy” approach would allow them to focus on the generative AI components for content analysis, processing and a UI rather than investing time and resources into building an internal platform for data collection from the ground up. The decision to use NewsCatcher significantly accelerated the project timeline and ensured a higher data quality from the outset.
Adopting the Solution
“We were looking for a news API that covered a significant amount of Caribbean media.”
To build the tool, the team partnered with NewsCatcher, which provided access to over 120 media outlets across the Caribbean. The API allowed them to extract and process news articles in multiple languages, including English, Spanish, French, and Dutch. Using LLMs, they categorized articles into ten predefined gender-related topics, such as gender-based violence, gender in the workplace and gender sexuality.
“I was impressed by the scope and how clean NewsCatcher’s data is.”
These articles were then summarized, tagged, and integrated into a chatbot interface, enabling users to query data on specific gender issues, such as workplace equality in Jamaica. With positive initial results, there are plans to further refine the tool and make it accessible for researchers, policymakers, and journalists.
Conclusion
The Gender Monitor project at the Mona School of Business & Management is an example of how NewsCatcher can be used to address region-specific issues. By focusing on gender narratives in the Caribbean, the team is creating a tool that will be able to provide valuable insights for researchers, journalists, and policymakers. The collaboration with NewsCatcher as a data source has been central to the project’s progress, allowing for the categorization and analysis of media articles across multiple languages.
Next, the team plans to explore Reddit as a complementary data source to gain insights from informal and community-driven discussions on gender issues. With NewsCatcher offering Reddit as part of its capabilities, this addition could enhance the project’s ability to capture diverse perspectives across the region.