Pangram Expands Suite to Include AI Detection for Images and Video

On Tuesday, Pangram—the company known for one of the most broadly used detectors of AI‑generated text—revealed that its system will now evaluate visual content, adding support for both videos and images.

Initially, the firm rose to visibility by providing a complimentary, web‑based utility that examines text for hallmarks of large language models. After its debut, the tool has been embedded in educational systems, publishing processes and content‑moderation workflows, assisting users in identifying material that might originate from AI.

With the newest release, Pangram reports that its algorithmic engine has been taught using a varied collection of synthetic visual datasets, allowing it to detect artifacts and signatures linked to generative image and video models. This step mirrors rising worries among educators, journalists and social‑media managers regarding how readily deep‑fake videos and AI‑generated graphics can spread.

Stakeholders will probably see the broadened capability as a sensible reaction to the expanding range of AI‑generated material. Although the solution may help platforms uphold authenticity rules, specialists caution that detection tools can generate false positives, particularly with stylized art or low‑resolution video. Pangram admits these issues and states it will keep improving its models using feedback from users.

Looking forward, the firm intends to launch an API enabling third‑party services to integrate the visual detection feature straight into their products. As the AI‑driven content creation arena advances, solutions such as Pangram’s could become a routine element of the digital‑media toolkit, assisting in preserving transparency and trust online.