Pangram Adds Video and Image AI Detection to Its Suite

Pangram, the firm behind one of the most widely adopted AI‑generated text detectors, announced on Tuesday that its technology will now assess visual media, extending its reach to videos and images.

The company first gained prominence for offering a free, browser‑based tool that scans written content for patterns typical of large language models. Since its launch, the service has been integrated into educational platforms, publishing workflows and content‑moderation pipelines, helping users flag material that may have been produced by AI.

In its latest update, Pangram says the algorithmic engine has been trained on a diverse set of synthetic visual datasets, enabling it to spot artifacts and signatures associated with generative image and video models. The move reflects growing concerns among educators, journalists and social‑media operators about the ease with which deep‑fake videos and AI‑created graphics can be disseminated.

Stakeholders are likely to view the expansion as a practical response to the widening scope of AI‑generated content. While the tool could aid platforms in enforcing authenticity policies, experts warn that detection systems can produce false positives, especially when applied to stylized artwork or low‑resolution footage. Pangram acknowledges these challenges and says it will continue to refine its models based on user feedback.

Looking ahead, the company plans to roll out an API that allows third‑party services to embed the visual detection capability directly into their own products. As the AI content creation landscape evolves, tools like Pangram’s may become a standard part of the digital‑media toolbox, helping to maintain transparency and trust across the internet.