Artificial intelligence (AI) is increasingly being utilized to assist journalists in finding newsworthy narratives and reliable sources. Researchers from the USC Information Sciences Institute are currently developing a source-recommendation engine that aims to suggest references for journalists based on a given text or topic. The software application will analyze the text and cross-reference it against a database of potential interviewees, experts, and informational resources. This tool will not only suggest relevant sources but also provide contact details, areas of expertise, and previous work of the sources.
Alexander Spangher, a computer science Ph.D. student at USC Viterbi and a former data scientist at the New York Times, is leading the development of this source-recommendation system. Having observed the challenges faced by journalists in traditional newsrooms, Spangher is motivated to provide helpful resources and tools for reporters. He acknowledges the increasing pressure on journalists in the face of news deserts and closures of newspapers, and sees AI as a potential solution to assist and support them in their work.
To create an AI model that can suggest sources, the researchers undertook a comprehensive study of how human journalists currently use sources in writing news articles. They gathered a dataset of sentences from over a thousand news articles and annotated the source of the information and the sourcing category. This dataset was used to train language models (LM), which are AI frameworks that process and understand human language. The trained LMs were able to detect source attributions with 83% accuracy. With these models, the researchers annotated approximately 10,000 news articles to gain further insights into how journalists use sources.
The AI models revealed that, on average, about half of the information in news articles came from sources. Each article typically had one to two major sources that contribute a significant portion of the information, along with two to eight minor sources that contribute less. The AI also found that the first and last sentences of an article were most likely to be sourced. This information is crucial for the development of the source-recommendation engine, as it helps determine when and how journalists currently use sources.
The researchers wanted to test if the AI algorithm could detect when a source was missing. A missing source could indicate incomplete information, and if AI can recognize this, it can recommend a particular expert to provide the missing piece. Analyzing 40,000 articles with randomly removed sources, the AI models successfully identified the absence of major sources but struggled with the detection of minor ones. Though less crucial to a story, minor sources can provide valuable recommendations and additional perspectives, which the AI system aims to recognize and recommend in the future.
In addition to aiding journalists in finding sources, the tool aims to promote diversity in journalism by recommending diverse voices beyond a journalist’s usual network. This can reduce reliance on familiar sources and bring fresh perspectives into reporting. However, the researchers acknowledge that AI systems can be prone to bias if not designed appropriately. To ensure diversity in source databases, it is essential to include representation from a wide range of demographics, disciplines, and perspectives.
Jonathan May, a research associate professor of computer science at USC Viterbi and ISI lead researcher, envisions a future where the source-recommendation engine will streamline the reporting process and make journalists more efficient. By assisting in the selection of sources, the tool can help journalists focus on the creative aspects of their work.
The research team plans to collaborate with journalists to gather feedback and further improve the tool. By incorporating input from journalists and understanding their needs and viewpoints, they aim to develop a solution that adapts to the requirements of local journalism. The successful implementation of AI technology in newsrooms has the potential to revolutionize the way reporters find sources and tell compelling stories.
Ravina Pandya, Content Writer, has a strong foothold in the market research industry. She specializes in writing well-researched articles from different industries, including food and beverages, information and technology, healthcare, chemical and materials, etc. With an MBA in E-commerce, she has an expertise in SEO-optimized content that resonates with industry professionals.