Pangram Expands AI Detection Technology to Identify AI-Generated Images and Videos
The AI detection startup is expanding beyond written content with new tools capable of identifying artificial intelligence-generated and edited images and videos as concerns about digital misinformation continue to grow.
Pangram, a New York-based artificial intelligence detection company, has introduced new technology designed to identify AI-generated images and videos. Announced on November 4, 2026, the expansion builds on the company’s existing reputation for detecting AI-written content. The new tools allow users to upload videos for analysis and determine whether artificial intelligence was used to create or modify the footage. The development comes as increasingly realistic AI-generated media makes it harder to distinguish authentic content from synthetic material.
One of the most notable features of the new video detection system is its ability to distinguish between completely AI-generated videos and footage that has only been partially modified using artificial intelligence. It also provides a visual heatmap highlighting specific sections of a video that may have been generated or altered by AI. According to Pangram, its internal testing showed that the video detection system could identify AI-modified content with more than 98 per cent accuracy.
The company has also upgraded its image detection technology through Pangram Image 1.0, which is now available through its Chrome extension. The extension can automatically scan images as users browse the internet and flag pictures suspected of being AI-generated or modified. Pangram claims its image detector identified more than 99 per cent of AI-generated images in internal testing, with a false positive rate of just 0.27 per cent. However, the company acknowledges that detecting AI modifications remains challenging, particularly because modern smartphone cameras also use AI processing to enhance ordinary photographs.
The new image and video detection features are currently available to paying subscribers, while free users can access a limited number of image scans. Pangram’s expansion reflects the growing demand for digital verification tools as AI-generated photographs, videos and other content spread across social media and online platforms. However, experts caution that AI detection systems are not completely reliable, meaning their results should not be treated as definitive proof that content is artificial. The development could nevertheless make it easier for journalists, businesses and ordinary internet users to investigate suspicious digital media.
