Free Fake Image Detector
Upload a photo and find out whether AI made it or a person altered it. FauxLens shows what each check found, so you can weigh the answer. Free to try, no account, and the file is deleted after analysis.
Upload the image or video. It is deleted after the check.Drop a photo or video here, or click to upload. It is deleted after the check.
JPG PNG HEIC · MP4
Detects content from
- OPENAI
- MIDJOURNEY
- STABILITY
- GEMINI
- FIREFLY
- FLUX.1
- CANVA
- IDEOGRAM
Key facts about this image check
How it works- Checks the file’s content credentials and metadata, the invisible watermark OpenAI and Google put in images their AI tools make, and the pixels, with a detection model.
- Takes a photo or a video: JPG, PNG, WebP, HEIC, MP4 or MOV.
- Gives a verdict, a confidence score and the evidence behind it.
- Your file is deleted after the check.
- A real photo used in the wrong context still comes back clean. Pair the scan with a reverse image search.
- Free credits while available. More come in one-time packs, no subscription.
What a result looks like
Every result comes with a confidence score and the evidence behind it. A low score means inconclusive, not clean.
AI-Generated or Likely AI-Generated
The checks found signs that AI made the image. If the evidence says only part of it was changed, the rest may be a real photo.
Do not act on it. Ask for the original, or check the claim another way.
No AI Detected or Likely Authentic
None of the checks found signs of AI generation in the image.
Checking a photo someone sent you
Most fake images do not arrive as news. They arrive in a chat, a marketplace listing, a dating profile or a post that is being shared faster than anyone can check it. A seller sends a photo of an item that does not exist. A new match has a profile picture that looks a little too perfect. A friend forwards a dramatic picture from an event nobody else is covering. In each case the question is the same: did a camera take this, or did someone make or change it?
Upload the image above and FauxLens checks it in seconds, with no account. It reads the file's content credentials and metadata, looks for the invisible watermark OpenAI and Google embed in images their AI tools make, and scores the pixels with a specialist AI-image detection model. The result is a verdict with short evidence lines saying what was found, so you can see why the answer is what it is.
Two habits make the answer more useful. First, check the best copy you can get: a photo that has been screenshotted, cropped and forwarded several times has lost much of what a scan reads. Second, treat a low-confidence result as inconclusive rather than as a clean bill of health, and combine it with a reverse image search, which finds where a real photo was taken from.
Where fake photos show up
Fake photos turn up wherever a picture is used as proof: marketplace listings, dating profiles, insurance claims, job applications, legal disputes and viral news posts. The question is the same in each: did a camera take this? One scan answers it. Each case below links to a page with the checks that matter most there.
Read moreShow less
- Buying and selling: listings built on AI-generated or borrowed product photos are a common way to sell something that does not exist. If the seller cannot send a fresh photo of the item on request, that tells you as much as the scan does. See the fake photo checker.
- Dating and social profiles: AI-generated faces are now easy to make and hard to spot by eye. A profile photo that comes back as AI-generated is a strong reason to slow down before sharing anything personal or sending money. See the dating profile photo checker.
- Insurance and claims: claim photos of flood damage, vehicle collisions and fire scenes are among the images most often fabricated or reused. A scan at intake flags likely AI-made photos before they reach manual review.
- Hiring and identity: trust and safety teams check that profile headshots are genuine photographs, because AI-generated professional identities are used to apply for remote jobs under false names.
- Legal disputes: AI-generated photos of property damage, accident scenes and alleged incidents have appeared in real proceedings. A documented scan result is one part of showing that an image was checked. See legal evidence verification.
- News and viral posts: dramatic pictures spread fastest in the first hours of a story, before anyone has checked them. See verifying news photos.
In each case the requirement is identical: a fast, evidence-backed answer about whether a photograph captures something real.
For newsrooms
Newsrooms need a repeatable way to check an image before it runs. The workflow, the press standards behind it and the topics where fakes cluster are below, and verifying news photos covers it in more depth.
Read moreShow less
Why a newsroom needs a check
AI-generated photographs are submitted to editorial desks, circulated as breaking news and used in coordinated disinformation campaigns. The latest generators produce photorealistic output that can pass editorial review by eye, and studies of people judging AI images by eye find them only somewhat better than chance. A retraction usually comes after the image has already shaped how people see the event, which is why fact-checkers treat a forensic check as a first step. FauxLens checks an image in seconds and returns a verdict with short evidence lines (content credentials and metadata, the invisible watermark OpenAI and Google embed, and a score from a specialist detection model), so it fits inside a deadline.
A five-step verification workflow
- Receive: when an image arrives from a source, stringer, wire service or social media, log where it came from and when. Provenance matters if the image is later challenged.
- Scan: upload the image before any editorial discussion. Review the verdict, the confidence score, the evidence lines and the AI vendor named when a watermark or content credential is found. If the scan flags it, mark it unverified and go to step 5.
- Reverse image search: run it through Google Images and TinEye to check prior publication. A brand-new AI image has no search results, which makes the forensic check your main tool for images appearing for the first time.
- Metadata audit: real press photography from professional cameras carries camera model, lens, exposure and often location data that is internally consistent. AI-generated images often carry none, or generic placeholder metadata.
- Escalate: if steps 2, 3 or 4 raise a signal, take it to your photo editor and ask the source for the original RAW file. AI generators do not produce RAW files, so a source who cannot provide an original leaves the image unauthenticated.
FauxLens is also available through an API for newsrooms that want the check inside their CMS or wire intake.
Press standards and AI images
Press bodies and major outlets share one principle: a photograph presented as documentary must show a real scene. The Associated Press, for example, says in its standards around generative AI that it does not alter any elements of its photos, video or audio, and does not allow generative AI to add or subtract any elements. Policies like these assume a verification step without prescribing one. Logging a FauxLens result in an image's editorial record is one practical way to show it was done, alongside reverse image search and contact with the source.
Where AI fakes cluster
AI-generated images cluster where public anxiety is high and authentic coverage is slow to arrive.
- Conflict zones: AI-generated photographs of alleged atrocities, casualties and military actions tend to appear in the first hours of an escalation, spreading through messaging apps before reporters on the ground can provide real coverage.
- Elections and politics: fabricated images of candidates and officials circulate most heavily in the days before major votes.
- Natural disasters: real photography is often unavailable at first and emotional urgency pushes people to share quickly, so AI flood, fire and earthquake imagery regularly appears presented as citizen journalism.
- Smaller outlets: with fewer photo staff and pressure to publish fast, they benefit most from a check that takes seconds and needs no account to start.
Is that photo or video real?
Related checks
Frequently asked questions
Yes. Save the photo from the chat or listing and upload it above. FauxLens returns a verdict with the evidence behind it, and the file is deleted after the analysis. Use the best copy you have: screenshots of screenshots and heavily compressed forwards hide much of what the scan reads.
It is free to try, with no account and no card. More scans come in one-time packs with no subscription.
They serve complementary purposes. Reverse image search finds where an image previously appeared online, but it cannot detect a brand-new AI-generated image that has never been posted before. FauxLens analyzes the image itself mathematically, detecting synthetic origin regardless of whether the image exists elsewhere online. Use both in combination for the strongest verification.
FauxLens is primarily optimized for AI-generated images. Edits made with AI tools such as Generative Fill come back as AI-generated, with a note that only part of the image was changed. Ordinary retouching without AI is not flagged. It does not show where an edit is, and careful copy-paste compositing can pass unnoticed.
FauxLens is built to flag AI-generated images from the generators most used to create fake news imagery, including Midjourney, DALL-E, Stable Diffusion, Flux, and Adobe Firefly. No detector catches every image, and small or heavily compressed copies are harder, so treat a low-confidence result as inconclusive. The evidence lines in the report give editors the basis for their verification decision.
The professional workflow combines reverse image search (to find prior publication history) with forensic pixel analysis (to detect AI generation or manipulation). FauxLens handles the forensic layer: upload any image and receive a verdict with a confidence score and evidence lines drawn from content credentials, metadata, the watermark OpenAI and Google embed in their AI images, and a specialist AI-image detection model.
Yes. The FauxLens API provides programmatic access for integrating image verification into CMS workflows, wire service intake, or content moderation pipelines. Contact support@fauxlens.com with your use case and volume requirements.
The National Press Photographers Association (NPPA) Code of Ethics prohibits altering the content of a photograph in any way that deceives the viewer. AI-generated images presented as documentary photography fall squarely under that principle. Disclosing any AI involvement and adding a forensic check at intake are the practical ways newsrooms apply it.
A practical newsroom AI policy covers four areas: intake verification (every incoming image is scanned with FauxLens before editorial use), escalation thresholds (confidence scores above 70% trigger photo editor review), documentation requirements (forensic scan results are logged in the image record), and disclosure standards (synthetic or AI-assisted imagery used editorially is labeled for readers). This policy can be implemented immediately using FauxLens as the forensic tool.
A forensic evidence chain in journalism refers to the documented sequence of verification steps that establish an image's authenticity. It includes: the provenance record (who provided the image, when, and from where), the forensic scan result (FauxLens verdict, confidence score and evidence lines), the reverse image search result, the metadata audit, and any source confirmation. Maintaining this chain protects the publication if an image's authenticity is later challenged.
Yes. Wire service images can and should be verified, particularly when they originate from unverified local sources that the wire service is redistributing. AP and Getty both have editorial standards prohibiting AI-generated content. But images from stringers and local contributors pass through multiple hands. Running a FauxLens check on any image whose provenance is not fully documented is sound editorial practice.
A FauxLens flag above 70% confidence is a signal to investigate, not automatically kill the image. Contact the source and request the original uncompressed RAW or TIFF file: AI generators do not produce RAW files, so this request is often determinative. Ask for camera make, model, and location metadata consistent with the scene. If the source cannot provide these, document that you could not authenticate the image and either do not publish it or publish without the photograph. Never publish an unresolved flagged image as documentary evidence.
Sources
Updated
Keep exploring
Guides
- The Journalist's Toolkit: How to Verify Any Photo Is Real in 2026AI-generated images are flooding newsrooms. Here is the step-by-step verification workflow used by digital forensics professionals and investigative journalists to separate real from synthetic.
- 5 Visual Tells: How to Spot an AI Image Without ToolsBefore you run a scan, check these 5 common mistakes AI models make. From topology errors in hands to subsurface scattering fails in skin.
- The Science of Deception: How AI Detection WorksA detailed guide to the digital forensics of AI image detection, from Error Level Analysis (ELA) and JPEG compression artifacts to Photo Response Non-Uniformity (PRNU), and which checks a FauxLens scan actually runs.
- Romance Scam Warning Signs: A Visual Checklist for 2026AI has made romance scammers more convincing than ever. FBI data shows over $1.3 billion lost annually. Here is the visual checklist every person on a dating platform needs before trusting a profile.
More tools
Detect while browsing: try the Chrome extension
Right-click any image · No account required