Case studies
Four real-world examples where FactNinja helped expose manipulation or unethical content.
⚠️ These cases are illustrative — names, figures, and context are anonymized or fictional for demonstration purposes.
Viral image of a "document from a protest"
A news editor came across a viral photo of a supposed "document" from a street demonstration. The post's author claimed it was official material from a political party.
Result: Within 30 seconds, FactNinja revealed that the typography and logo did not match any public material from the party. Reverse search showed the image had first appeared on an anonymous 4chan profile.
The newsroom did not publish the story and instead wrote about the disinformation campaign.
Teaching media literacy at a high school
A civics teacher uses FactNinja in classes with second-year high school students. Every week, they analyze 2–3 real social media posts and discuss manipulation techniques.
Result: Students learned to recognize emotional appeals, false dichotomies, and manipulative photo composition. In an end-of-semester survey, 82% of students said they now think more critically about content on social media.
The teacher shares analyzed example sets with colleagues via public groups.
Independent ethics review of an election campaign
A civic association monitoring election campaigns filed a lawsuit against a political party for suspected xenophobic and racist content in its campaign visuals. They needed a structured, reproducible analysis so the court could assess whether the visuals crossed the line of free speech.
Result: Without human intervention, FactNinja evaluated 47 campaign visuals and identified specific rhetorical and visual techniques in 23 of them (dehumanization of a minority group, false associations, distorted statistics). The output was used by attorneys as independent expert evidence.
The system works equally well for evaluating your own campaign before publication.
Mapping a disinformation campaign on Telegram
A team of academic researchers studied a network of Telegram channels that coordinated the sharing of nearly identical visuals with edited captions. They needed to systematically process hundreds of images and document their manipulative patterns.
Result: They uploaded 312 images, sorted them into 8 categories, and created a public group with the analyses. Based on the output, they published a peer-reviewed study on coordinated information influence during regional elections. The data is traceable and consistent.
The group is publicly shared and also serves as teaching material.