Image Fact Checker: How to Check If a Photo Is Real
BSmeter doesn't analyze images. It can't tell you whether a photo is AI-generated, Photoshopped, or a deepfake. This guide covers the checks that work (reverse image search, Content Credentials, tracing the original source) and where BSmeter helps: checking the claims in the videos and podcasts that spread those images.
How to Fact-Check an Image in 3 Steps
Checks anyone can run, no special tools required
Reverse-Search the Image
Run the image, or a screenshot of a video frame, through a reverse image search such as Google Lens, TinEye, or Bing Visual Search. Earlier copies can show where and when it first appeared, and whether it's being reused from a different event.
Check Provenance and Labels
Some images carry Content Credentials (the C2PA standard) that record how they were made or edited, and you can inspect them with a Content Credentials verify tool. Many platforms also label AI-generated media. Most images don't carry credentials yet, and metadata is easy to strip, so a missing label proves nothing.
Check the Claim Around It
Misleading images usually come with a claim attached: a date, a place, a number. Check it against reputable reporting and official channels. If the image appears in a video or podcast, BSmeter can check the spoken claims. Paste the link to get each claim scored 0-100, with evidence sources for the weakest one (the 3 weakest on Basic and Premium).
What BSmeter Checks — and What It Doesn't
BSmeter analyzes the claims in videos and podcasts. It does not analyze images.
No AI-Image Detection
BSmeter does not tell you whether an image was made with DALL-E, Midjourney, Stable Diffusion, or any other generator. Use the checks on this page instead. Dedicated AI-image detectors exist, but their results are probabilities, not proof.
No Photo-Manipulation Forensics
BSmeter doesn't inspect pixels or metadata, doesn't look for cloning, splicing, or Photoshop edits, and doesn't analyze screenshots. It also has no image upload. It works from video, podcast, and audio-file links.
Describes On-Screen Frames for Context
When BSmeter analyzes a video, a vision pass can describe what's on screen so the spoken claims are read in context. It describes the frames. It does not judge whether they are real, edited, or AI-generated.
0-100 Score for Each Claim
BSmeter transcribes a YouTube, TikTok, Instagram Reel, podcast, or audio-file link, pulls out the verifiable claims, and gives each one a 0-100 credibility score.
Evidence Search for the Weakest Claim
For the weakest claim (the 3 weakest on Basic and Premium), BSmeter searches for evidence sources you can open and read. Sources can be imperfect and results are AI-generated, so judge them for yourself.
Framing Moves & Fallacies
BSmeter names rhetorical moves such as Missing Context, Cherry-Picked, Anonymous Authority, and Confidence Mismatch, plus logical fallacies. Those are the same tricks behind many out-of-context images, applied here to what's said in a video.
How to Spot Fake Images
Common warning signs that an image may be AI-generated, manipulated, or misleading. None of them is proof on its own.
Unnatural Hands, Fingers, and Anatomy
AI image generators have struggled with human anatomy: extra or merged fingers, mismatched eyes, teeth that look like a single block, or limbs bending at impossible angles. Newer generators make fewer of these mistakes, so a clean image isn't proof that it's real.
Lighting and Shadow Inconsistencies
Manipulated and AI-generated images often have lighting that doesn't match the scene: shadows falling in impossible directions, multiple light sources that contradict each other, reflections that don't match the reflected object, or lighting that differs between the subject and background.
Out-of-Context and Recycled Images
One of the most common image misinformation tactics: sharing a real photo from a different event, date, or location as if it's related to a current news story. An old war photo becomes 'today's protest', or a past hurricane image gets shared as 'tonight's storm'. Reverse image search is the best way to catch this.
Missing or Mismatched Metadata
Photos straight from a camera or phone usually carry EXIF data such as the date, device, and sometimes location, which you can inspect with a metadata viewer. Many social platforms strip this data on upload, so missing metadata isn't suspicious on its own. A timestamp that contradicts the claimed event is worth a closer look.
Frequently Asked Questions
Straight answers about image fact-checking and what BSmeter does
Does BSmeter detect AI-generated images?▼
No. BSmeter doesn't analyze images and can't tell you whether one was AI-generated. To check an image yourself, reverse-search it, look for Content Credentials or a platform AI label, and trace it to its original source. BSmeter's job is claims: paste a video or podcast link and it scores each verifiable claim 0-100.
Can BSmeter tell if a photo has been Photoshopped?▼
No. BSmeter doesn't inspect pixels or metadata. Signs worth checking yourself include inconsistent shadows and lighting, warped edges or backgrounds near the subject, and mismatched reflections. Skilled edits can leave no visible trace, though, so finding the original through reverse image search is usually more reliable.
Does BSmeter check images in videos?▼
Only for context. When BSmeter analyzes a video, a vision pass can describe what's on screen so the spoken claims are read in context. It does not check frames for manipulation, face swaps, or AI generation. What it checks is what's said: each verifiable claim gets a 0-100 credibility score, and it searches for evidence sources for the weakest claim (the 3 weakest on Basic and Premium).
Is there a free version for image fact-checking?▼
BSmeter doesn't check images, so there's no image plan. Free reverse image search tools like Google Lens and TinEye are a good place to start. For videos and podcasts, BSmeter's free plan gives 3 analyses a day for videos up to 30 minutes with a free account; Basic is $3.99/month (50 a day, up to 60 minutes) and Premium is $8.99/month (videos up to 2 hours).
How is BSmeter different from reverse image search tools?▼
They do different jobs. Reverse image search shows where an image appears online, which often reveals its original date and context, so use it for images. BSmeter doesn't search or analyze images. It checks the claims made in YouTube videos, TikToks, Instagram Reels, podcasts, and audio files.
Can BSmeter detect deepfake images of people?▼
No. BSmeter doesn't detect face swaps or deepfakes. Look for provenance (Content Credentials, the original uploader, the person's official channels), reverse-search the image, and be wary of emotionally charged images with no traceable source. If a deepfake video makes claims, BSmeter can check those claims.
What about screenshots — can those be faked?▼
Yes, easily. Fake screenshots of posts, headlines, and messages are common. BSmeter doesn't analyze screenshots. The reliable check is to find the original: search the account or outlet for the post or headline, and check archive services such as the Wayback Machine.
How accurate is BSmeter's image analysis?▼
BSmeter doesn't do image analysis, so there's no accuracy figure to give. Its claim analysis is AI-generated and can be wrong, and evidence sources can be imperfect. Use the scores as a starting point and open the sources yourself.