Is This Image on X AI?
Grok can make a picture and post it from the same app, viral images are reposted without credit within minutes, and a checkmark on X is something an account pays for. Here is how to check an image you found on X, starting with the detector.
Drag & drop, paste, or click to upload
PNG · JPG · WEBP · HEIC — your image is analyzed privately and deleted
This is an estimate, not proof — the detector can be wrong, producing both false positives and false negatives. Don't rely on it as your only source; verify anything important. wedetect.ai is not liable for decisions made from these results. See our Terms.
1. Save the largest copy X keeps
Open the image so it fills the screen and save it from there, not from the timeline preview. In a desktop browser, open the image in its own tab: the address ends in name=small or name=medium, and changing that to name=orig loads the largest version X stored. In the app, screenshot the opened image and crop away the buttons.
X removes the metadata and any Content Credentials from every upload, so a copy saved from X has nothing for the metadata checker to read, whether it is real or generated. The check on the pixels themselves still works.
2. Run the AI detector
Drop the saved image into the checker above. It returns the probability that a generator made the image, not a yes or no. A high score is a reason to look harder before you repost, and a low one is not a guarantee. For anything important, pair it with the steps below.
3. Rule Grok in or out
Grok, the assistant built into X, generates images on request, so on X it is the first generator to consider. What its pictures tend to look like, and why they carry no provenance data at all, is covered in how to tell if an image is Grok.
Many people reply "@grok is this real?" under a viral picture. Grok answers like any chatbot: from what it can find and infer, without examining the file, and it can be confidently wrong. Treat its reply as one more opinion in the thread, not as a check.
4. Read the context X gives you
- Community Notes. Notes are written and rated by contributors who usually disagree with each other. A note can be attached to the image itself, so the same note follows every post that reuses it. A note is strong evidence; no note means nobody has written one yet.
- Replies and quote posts. Someone has often found the source already. Sort through them for a link to an earlier post, a photographer or a news outlet.
- The account. A checkmark on X shows a paid subscription, not a confirmed identity. Look at the join date, what else the account posts, and whether it credits anyone. Accounts that post a stream of viral images with no sources are farming engagement.
- Labels. X can label a post as manipulated or synthetic under its own media policy, but most AI images carry no label at all, so a missing label tells you nothing.
5. Find the first post
Search the image with Google Lens, TinEye or Bing Visual Search to see where else it appears, then search X itself for the earliest posts using the subject and a date range. A "breaking" picture that turns up months earlier, or on an AI art account, settles the question on its own. Where did this image come from? walks through the full search.
Putting it together
| A viral picture of a news event | Reverse search first; old photos and AI art are reused as "breaking". Then the detector. |
|---|---|
| An image with a Community Note | Read the note and its sources, then check the image yourself; notes can lag or miss. |
| A "@grok is this real" reply | An opinion, not a test. Run the detector and the search. |
| A profile picture of an account in your replies | Detector plus a reverse search of the picture; see the fake profile picture detector. |
Frequently asked questions
How can I tell if an image on X is AI?
Save the largest copy of the image, run it through an AI image detector for a probability, check whether Grok could have made it, read any Community Note, and reverse search it to find the first post. Two checks that agree are worth more than one score.
Does X label AI-generated images?
Sometimes. X can label posts under its policy on synthetic and manipulated media, and Community Notes can add context, but most AI images on X carry no label. No label does not mean the image is real.
Is asking "@grok is this real" a reliable check?
No. Grok answers from what it can find and infer, without examining the file, and like any chatbot it can be confidently wrong. Use a detector and a reverse image search instead.
Why is there no metadata on images saved from X?
X re-encodes uploads and removes EXIF data, location and Content Credentials, so a saved copy never carries the original capture data. Metadata tools show nothing either way; the pixel check is the one that still works.
Can a Community Note on an image be wrong?
It can be late, incomplete or occasionally wrong, because contributors write it from what they can find. Read the sources a note links to, and treat it as strong evidence rather than the final word.
Related: how to tell if an image is Grok · WhatsApp photo check · Instagram photo AI check · Facebook photo AI check · where did this image come from?.