Media Matters for Democracy (MMfD), a digital rights and media literacy nonprofit based in Pakistan, is working on a new AI-powered tool to assist journalists.
The project, which is in early stages of development, will serve as a convenient “research assistant” for reporters, according to details.
The initiative, called the “Smart News Assistant”, will enhance the capacity of journalists specifically in research, enabling them to access the required information faster from old documents and archives.
“AI has the potential to either deepen the crisis journalism is already facing or to become a powerful force for renewal and empowerment,” says Asad Baig, Executive Director at MMfD. “At Media Matters for Democracy, we are focused on the latter i.e., working directly with journalists and newsrooms to infuse AI into journalism processes in ways that strengthen reporting, not replace it.”
The Smart News Assistant will also ensure accuracy by digitally investigating pieces of information against reliable sources to help uncover patterns and connections that might be missed during traditional research.
The project will deploy open-sourced Large Language Models (LLMs) like Llama 3, which is an artificial intelligence system with the ability to comprehend questions and fetch related information.
“We believe the real promise of AI lies in how it can be used to automate repetitive tasks, analyse massive datasets, surface hidden patterns, and support investigative work that would otherwise be overwhelming given the resource constraints so many Pakistani newsrooms face,” Baig says.
Another key component of the project is a “central knowledge base”, which will be accessible to newsrooms. It will include their archives and information from other sources such as the Pakistan Bureau of Statistics (PBS).
On the other hand, a “smart indexing” technology will digitally capture and highlight key facts and connections from its archive. This will enable faster retrieval of relevant information.
At the core of the system is an integration using Langchain tools, which link the digital archive to the AI’s analytical engine. When a journalist submits a query, the system will scan the archive, extract relevant data, and deliver it to the AI for real-time analysis.
To streamline the process, MMfD is also developing a user-friendly chat-style interface that will allow journalists to type questions and receive immediate responses.
“If we empower journalists to use AI as a tool for efficiency, verification, storytelling, and audience engagement, while maintaining standards, it can help journalism become faster, deeper, and more resilient,” Baig adds.
