Pangram extends AI detection to video and images, with paid tiers for full access
The tools arrive after an early image model wrongly flagged genuine iPhone 18 Pro photos, which delayed its release.
Pangram, known for a widely used AI text detector, has added tools that flag AI-generated video and images. Subscribers can upload clips to receive an assessment that separates partially AI-edited footage from fully synthetic footage, along with a heatmap showing which parts are likely AI. The company reports its video model exceeded 98 percent accuracy in one internal test and outperformed other detectors. Its image model, Pangram Image 1.0, is now available through a Chrome extension that scans images while browsing, and it reports above 99 percent detection with a 0.27 percent false positive rate, though results below 512 x 512 pixels carry a reduced-confidence warning. The company said it delayed the image model's release because early versions flagged genuine photos from iPhone 18 Pro cameras, a problem tied to on-device AI processing. The new features require a paid Pangram subscription, while free accounts can check a few images.
Why it matters
For readers who rely on photos and video as evidence, this is a sign that AI detection is spreading beyond text, but the company's own figures come from internal tests and have not been independently checked. The iPhone 18 Pro episode shows how on-device AI processing can make real photos look synthetic to a detector, so a flag is not proof on its own. Free accounts can check only a handful of images, so casual checking stays limited unless people pay.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.