Is it safe to upload your photos to an AI detector?
Most detectors are an upload. What that means for a claim photo, a passport or client work under NDA, and the questions worth asking before you paste anything in.
Read guide
Guides
Understand the evidence.
Clear, practical writing about image forensics — what automated signals can tell you, what they cannot, and how to verify responsibly.
Why AI still gets hands wrong, and when it does not
Hands were the most quoted tell in AI imagery and are now mostly fixed. What made them hard, what solved it, and what the episode says about every other visual tell.
Read guideBuilding an image check into your own pipeline
There is no API here, and this explains what that means for automation, what a browser tool can and cannot do for you, and what the alternatives actually cost.
Read guideFinding AI images inside PDFs and documents
Reports, decks and claim bundles carry images that nobody checks. How to extract them without losing quality, and which ones are worth the effort.
Read guideWhy screenshots break AI detection
A screenshot replaces the evidence a detector reads with evidence about your own screen. What survives, what does not, and what to do when a screenshot is all you have.
Read guideAI-upscaled vs AI-generated
Upscaling invents pixels that were never captured, and scores like generation because in a narrow sense it is. How to tell an enlarged photograph from an invented one.
Read guideSynthID and invisible watermarks
Some generators embed a signal in the pixels themselves. What survives editing, what does not, and why a watermark check and a pixel check answer different questions.
Read guideWriting an AI-image policy for your team
A short, usable policy beats a long one nobody reads. The four decisions that matter, the wording that works, and the clause your suppliers need.
Read guideThe EU AI Act and image labelling
Transparency obligations for synthetic imagery, who they fall on, and what a business outside the EU should do about them. A plain summary, not legal advice.
Read guidePhoto contests and AI disqualification
Competitions now check entries, and the rules turn on editing limits rather than on generation. What organisers should require, and what entrants should keep.
Read guideStock photo sites and AI submissions
Libraries went from banning generated work to selling it. What each side of the marketplace needs to check now, and why undisclosed submissions are the real problem.
Read guideDo AI images hurt your SEO?
Search guidance is about usefulness rather than production method. What actually costs rankings, what does not, and where image provenance is starting to matter.
Read guideWho owns an AI-generated image?
Copyright in generated pictures is unsettled and differs by country. What is broadly agreed, what is contested, and why detection matters to the question at all.
Read guideAI images in election misinformation
Most misleading political images are real photographs with false captions. Where detection helps, where it actively misleads, and what to do in the hours that matter.
Read guideCharity appeals using AI images
Generated imagery in fundraising raises money and destroys trust when it surfaces. How to check an appeal you are asked to give to, and what charities should do instead.
Read guideAI-generated fake IDs: what gets caught
Wholly fabricated documents are the easy case. The one that passes most checks is a genuine document with a single field altered, and it needs a different kind of look.
Read guideAI images in crypto and investment scams
Fake trading screenshots, invented offices and generated founders. Which parts of an investment pitch a pixel check reaches, and which need a register lookup instead.
Read guideAI images in Google Image results
Search results now mix photographs and generated pictures without much to separate them. How to filter, what the labels mean, and how to check before you reuse anything.
Read guideRomance scams and AI photos
The photograph is rarely the weakest link, and checking it rarely saves anyone. What actually identifies a romance scam, and where an image check does help.
Read guideAI product images on Amazon and Etsy
Generated product photography sells items that do not exist as pictured. What to look for as a buyer, and what sellers need to know before their listing gets removed.
Read guideAre these Airbnb photos real?
Virtual staging, replaced skies and rooms that were never that wide. What a pixel check finds in a holiday rental listing, and the four questions it cannot answer.
Read guideAI photos on LinkedIn profiles
Generated headshots are ordinary now and say nothing about a candidate. A fabricated professional identity is the real problem, and the photo is not how you catch it.
Read guideChecking images from Facebook Marketplace
Most fake listings use stolen photographs rather than generated ones. Here is the two-minute check that catches both, and the request that ends the question outright.
Read guideHow to check if a Facebook photo is AI
Facebook recompresses everything it touches, which changes what a detector sees. How to get a usable copy of the image, and how to read the result once you have it.
Read guideSpotting AI images on Instagram
Filters, presets and a platform built on retouching make Instagram the hardest place to read a detector score. What still works, and what to stop treating as evidence.
Read guideDeepfake vs AI-generated: not the same problem
One invents a person who does not exist. The other puts a real person somewhere they never were. They need different checks, and conflating them causes real mistakes.
Read guideHow to check if a profile picture is real
A pixel check answers one question about a profile photo. Four other checks answer the ones that actually matter, and together they take about two minutes.
Read guideChecking AI images on iPhone and Android
No app to install and no upload. How to check a picture from your camera roll, a message or a web page on a phone, and what phone processing does to the score.
Read guideHow to spot AI-generated faces
The visual tells everyone repeats are the ones that got fixed first. Here is what still separates a generated face from a photographed one, and what no longer does.
Read guideIs this photo AI? A 60-second check
A photograph landed in front of you and something feels off. Here is the fastest reliable way to decide, what to look at first, and when to stop guessing.
Read guideAI headshots in recruitment: where it matters
A generated profile photo is usually harmless. A generated identity document is not. How to tell which problem you have, and where checking crosses a legal line.
Read guideChecking images in bulk: a batch workflow
How to review twenty-five images at once, what to look for in a set rather than a single result, and why the outlier matters more than the highest score.
Read guideDiffusion vs GAN: why detectors handle them differently
Two ways of building an image, two different fingerprints. What each process leaves behind, why GANs were easier to catch, and what that predicts about what comes next.
Read guideA field guide to image generators and what they leave behind
What the pixel model can estimate, what the file can name, and why a detector that claims to identify the tool behind an image is overreaching.
Read guideFree vs paid AI image detectors: what you are paying for
Paid tiers buy reporting, history and API access. They rarely buy accuracy. Here is how the market actually prices this, and when paying is the right call.
Read guideCan ChatGPT tell if an image is AI generated?
It will give you a confident answer with nothing behind it. Why language models cannot do this, what they are actually doing, and what to use instead.
Read guideEvery way an AI image detector can be wrong
A complete failure taxonomy: what breaks, why, and what each failure looks like when it lands in front of you. Written to be read before you trust a score.
Read guideHow AI image detection actually works
What happens between dropping a file in and reading a score: the pixel model, the tiling pass, the file evidence, and how the three are combined.
Read guideHow to export and share an image analysis report
Three formats, three audiences. What goes in a client PDF, what belongs in a case file, and how to share a result without sharing the picture.
Read guideVerifying supplier and manufacturer photos
Sample photos decide bulk orders worth thousands. How to check what a supplier sends before a deposit moves, and what to require instead of trusting an image.
Read guideWhat a region heatmap shows that a score cannot
One number collapses three different situations into one answer. The tile view separates them, and the difference usually decides what you do next.
Read guideWhat happens to an image every time it is re-saved
Each save discards detail permanently. We ran one photograph through ten cycles and scored it at every step. The result explains most false positives.
Read guideC2PA and Content Credentials, explained plainly
What a signed manifest proves, what an absent one does not, and why the standard is stronger than pixel analysis and still cannot solve this on its own.
Read guideWhy the detector runs in your browser
A 40 MB model, WebAssembly and no upload endpoint. What that architecture buys, what it costs, and why the trade is worth making.
Read guideWhy EXIF metadata cannot detect AI images
Metadata is stripped by almost every platform on upload, which means it fails on precisely the images people need to check. What survives, and what to use instead.
Read guideAI images and academic integrity
Why a detector score must not decide a misconduct hearing, what institutions should do instead, and how students can evidence their own work.
Read guideCan AI image detectors be fooled?
Yes, and knowing how is what makes a result readable. The categories of attack, why some work, and what a laundered image looks like when it lands.
Read guideChecking real estate listing photos
Virtual staging is legitimate and disclosed. Generated rooms are not. How to tell the difference before you pay a holding deposit on a property that does not exist.
Read guideFake receipts and expense fraud
Generated receipts now pass a glance. A checklist for finance teams, why altered totals leave a local signal, and where the check has to stop.
Read guideModerating AI-generated images at scale
Triage thresholds, batch workflows and why an automatic ban on a score is the wrong policy. A working guide for trust and safety teams.
Read guideWhere an AI check fits in newsroom verification
Detection sits after sourcing and before publication, never instead of either. A working sequence for picture desks, and what to do when the result is uncertain.
Read guideHow to prove your artwork is not AI
Designers and photographers are being asked to prove authorship. A detector score will not do it. Here is the evidence that does, and how to build it as you work.
Read guideSpotting AI photos on dating profiles
Romance scams now run on generated faces that pass a glance. What still gives them away, what to ask for, and the checks that work before you meet anyone.
Read guideVerifying photos in an insurance claim
Generated damage photos are cheap to make and expensive to pay out. A workflow for claims handlers, and the point at which it must stop and go to an examiner.
Read guideWhy AI detectors disagree on the same image
Run one picture through four tools and get four answers. The reasons are structural, and knowing them tells you which result to believe.
Read guideAI image detector accuracy: what 91% actually means
Balanced accuracy, decision lines and the difference between a benchmark and your image. What the headline number covers, and what it quietly leaves out.
Read guideAI image detector vs reverse image search
They answer different questions. Reverse search finds where a picture has been; a detector estimates how it was made. Here is when to use each, and why order matters.
Read guideHow to check if a product photo is real before you buy
Generated product shots sell items nobody made. Here is the four-minute check that catches them, in the order that finds fraud fastest.
Read guideHow to tell if an image is AI generated
The visual tells that still work, the ones that stopped working, and how to check an image properly when looking at it is no longer enough.
Read guideWhy a real photo gets flagged as AI
False positives have six common causes, and most of them are fixable. Here is why a genuine photograph scores high, and how to re-test it properly.
Read guideHow to read an AI image detector score
What the number means, how the five bands work, when the region map overrules the headline score, and the point at which the honest answer is uncertain.
Read guide