In 2018, identifying a false news story meant doing almost everything manually. I would find a suspicious article, examine the claims it made, search original sources, cross-check the information, and then write an article explaining what was false, misleading, not verified, and unsupported.
It was time-consuming work, but the principle behind it was simple and remains fundamental to journalism: verification. The aim of the project called The truth online was to raise awareness among people in Kosovo that not everything they read online is based on facts, and not everything is true.
It was an interesting time in journalism in Kosovo, where most of the newspapers were going online, and people with no background in journalism were trying to imitate journalists who were just trying to claim ground in the online news marketplace.
As we explain in our training courses with the teachers in Kosovo, journalism is a cultural practice of telling verified and nonfiction stories that inform people about what separates journalism from advertisements and public relations. This practice for journalists usually starts by respecting the core journalism values, such as verification of information through multiple sources, reliance on experts, review of evidence, and a rigorous editorial review.
Eight years later, that principle has not changed. The information environment around it has.
The volume of digital content and the speed with which it can circulate online has increased enormously. False and misleading information can travel rapidly through news websites and social media accounts. At the same time, small organizations working on journalism and media education often operate with limited human resources. Manually checking every questionable article is simply impossible.
This was the problem that led me back to work I had started years earlier, but this time with a new question: Could artificial intelligence (AI) help us scale verification without replacing the journalistic principles behind it?
Manual verification to IS IT TRUE
That idea became IS IT TRUE, a project currently being developed by the Spiranca Institute in Kosovo. The project takes something that began as a manual process and explores how technology can help us work with a much larger amount of news content.
IS IT TRUE is being developed as an AI-supported tool that analyses news articles and provides an assessment of whether the content may be true or false. The tool is built around a dataset of articles that allows the system to recognize patterns in the content it analyses. It will be integrated into the Spiranca Institute’s website as a standalone page.
For us, however, the project is not only about producing a result. An important part of its development is making the assessment understandable to the user rather than simply presenting an unexplained conclusion. We do this by utilizing traditional journalism verification methods.
The intention is to create a tool that can support the work of identifying questionable information while making this process more accessible to people beyond professional fact-checking organizations. A central part of this work has been the dataset on which the tool is being developed.
The next step is to complete the development and testing of IS IT TRUE and integrate it into the Spiranca Institute’s website. From there, we want to continue improving the tool as we learn from its use and expand the work around verified journalism and media education.
Making verification more accessible
Once developed and tested, IS IT TRUE is intended to become publicly available through the Spiranca Institute website.
Because Spiranca Institute is a youth-centered organization, this project was also designed particularly with young people in mind. But the potential audience is broader and includes journalists, educators, students, and researchers.
For our organization, this is also important because the project brings together areas that we normally approach separately: Verified journalism, media education, and the responsible use of artificial intelligence.
The question remains the same
Eight years after I first began manually checking questionable news stories, the basic question has not changed: Is it true?
What has changed is the scale of the information environment and the technology available to navigate it. The next stage is to test, refine, and integrate the tool into Spiranca’s website, while continuing to examine its accuracy, limitations, and potential educational value.
The intention is for it to remain publicly accessible after the project ends and to improve as the dataset and our understanding of the technology develop. The aim of IS IT TRUE is to use the machines to navigate in environments where there are no clear gatekeepers.
While we believe that journalists alone cannot be responsible for every piece of information that is being spread online — people must be able to assess information on their own — it is up to journalists to safeguard the industry and our own professional boundaries.
