Built a first-pass AI video screening tool — looking for feedback on cautious report wording
I’m working on AI Video Detector, a web tool for people who see a suspicious social video and want a quick first-pass signal report before sharing it.
The main product constraint is that this is a sensitive category. I don’t want the tool to overclaim that it can prove a video is real or fake. The report uses likelihood, confidence, source/media access, visual artifact, motion, compression, and audio-video consistency signals where available.
I’d appreciate feedback on:
- Is the positioning clear enough for a non-technical user?
- Does the wording avoid overclaiming?
- What would you want to see before trusting a report enough to investigate further?
Trending on Indie Hackers
The part I'd pay attention to isn't whether the wording overclaims.
It's whether users interpret uncertainty the way you intend them to.
Two people can read the same cautious report and walk away with completely different levels of confidence.
That's the kind of thing that can be hard to notice early because the wording itself may look perfectly reasonable.