University of Washington researchers have launched PaperTok, a free tool that converts dense academic papers into 45-second accessible videos. The project aims to combat scientifically inaccurate AI-generated content by giving scientists direct control over how their complex findings are translated for short-form video platforms.
Translating dense research
The tool requires a researcher to upload a paper, which the system then analyzes to identify attention-grabbing hooks and relevant takeaways for a general audience. PaperTok generates a script with a narrative arc and produces an AI-narrated video that closes with a formal citation to establish credibility. Unlike other PDF-to-video converters, the system uses a multi-step process that requires human approval at each phase. Users can edit the output down to individual words before finalizing the clip.
Built by scientists, for scientists
Meziah Ruby Cristobal, a UW doctoral student in human centered design and engineering, co-led the development of the tool alongside fellow doctoral student Donghoon Shin and professor Gary Hsieh. The team of eight built PaperTok last summer after interviewing science communicators and gathering user feedback. Cristobal said short-form platforms like TikTok and YouTube Shorts have become primary venues for sharing scientific findings. The team presented their work this spring at the Association for Computing Machinery's Conference on Human Factors in Computing Systems.
Refining the output
Early users found value in seeing the system visualize abstract concepts, often using it as a brainstorming tool to find new ways to communicate their Research. "A lot of the researchers actually found huge value in seeing how the AI tries to visualize what they believe to be very abstract concepts," Cristobal said. Some users provided critical feedback, pointing to issues like generated nonsense text that made the videos feel too artificial.
The UW team is actively refining the platform and plans to let researchers incorporate their own charts and graphics into the final videos. While originally built for human-computer interaction papers, the tool has tested well on physics topics. The project highlights the growing role of AI for Science & Research, demonstrating how machine learning can translate complex data for public consumption. The tool is free to use, though researchers must provide a Gemini API key to cover the computational costs of video generation.
Why this matters for science and research professionals
Scientists and researchers face a growing problem where AI-generated summaries misrepresent their published work on social media. PaperTok offers a direct method to reclaim that narrative, allowing labs to produce accurate, citable video abstracts without requiring dedicated media teams or video editing skills. By keeping humans in the loop, the tool ensures that the final output reflects the actual science rather than an algorithmic approximation.
Your membership also unlocks:
