Why This Expansion Matters Right Now
If you have spent any time scrolling social feeds lately, you have probably noticed how hard it is to tell a real photo from one cooked up by a model in seconds. Pangram, the company known for its popular tool that checks whether written text was produced by AI, has now stepped into the visual world. Its new tools can examine videos and images and estimate whether artificial intelligence was used to create or edit them. For anyone who cares about what they are looking at online, that is genuinely exciting news.
The timing makes sense. AI-generated video has improved dramatically, and clips that once looked clumsy now pass casually as footage from a phone. Pangram says its video model is especially strong at spotting fully synthetic content, and in one internal test it reportedly flagged AI videos with more than 98 percent accuracy. Those are the kinds of numbers that make a difference when a viral clip could sway opinions, shape a purchase decision, or damage a brand overnight.
What the Tools Can Actually Do
The video checker is designed to go beyond a simple yes or no. Pangram says it can distinguish clips that are partially AI-edited from those that are entirely generated, which matters a lot because so much modern content is a mix of real footage and digital touch-ups. A heatmap feature also highlights which parts of a video look likely to be artificial, giving you a visual guide rather than just a verdict.
On the image side, the company has brought its research preview into a Chrome extension. The extension scans pictures as you browse and flags suspicious ones automatically. Pangram reports that its image model caught more than 99 percent of AI-generated pictures in testing, with a false positive rate of just 0.27 percent. The company also notes that very small images, below 512 by 512 pixels, come with a reduced confidence warning, and it admits that results can wobble on tricky cases.
The Fine Print Worth Knowing
No detector is perfect, and Pangram is refreshingly open about the challenges. Modern phone cameras, including iPhones, apply some AI processing to add detail to ordinary photos. That created a real headache during development, since the model initially flagged genuine camera shots as artificial. The team delayed its release to test on iPhone 18 Pro devices and fine-tune the system so it can tell ordinary computational photography apart from fully fabricated visuals. That kind of care is reassuring, because a detector that cries wolf at every selfie would quickly become useless.
Access works on a subscription model, though free accounts can check a handful of images before you need to pay. That makes it easy to test the waters before committing.
What This Means for Everyday Shoppers and Creators
For consumers, tools like this could soon become part of everyday online shopping. Think about fake product demo videos, doctored reviews with staged images, or influencer clips that were never filmed in the real world. A browser extension that quietly flags suspicious visuals can help you avoid costly mistakes before you click buy. Creators, meanwhile, can use the same technology to prove their work is authentic, which may become a selling point as audiences grow wary of synthetic content. As more people adopt verification tools like these, trust itself starts to become a feature worth paying for, and that is a shift every savvy buyer should be watching closely.
