Screenshots that look like real posts are often false. The evidence is checkable in minutes. A guide by Shayan Sardarizadeh for the Global Investigative Journalism Network lists the recipes. Fabricators use photo apps, generator sites, or an edit of the code on a real post. One fact anchors the topic: a screenshot is an image, not a clickable post.
The fabrications fail basic checks. The account timeline shows no such post. The archive holds no record. Fonts and dates betray the edit. X calls deceptively altered media misleading under its authenticity rules. The sections below cover the methods, the tells, one documented failure, and the checks that settle a doubtful image.
How are fake screenshots made?
The recipes repeat across formats. A maker retouches a real capture in an image editor. Another types invented text into a generator site. A third opens the browser inspector on a real post and rewrites it. Envista Forensics, a digital forensics firm, lists generator sites for fake text threads. Those sites let a user pick timestamps, carrier icons, and battery levels. Device settings offer a no-tools route. Change the phone clock, and the timestamp changes with it. Edit the contact name, and the capture invents a sender. The output is always a flat image. A real post never is.
Which visual tells give a fake away?
Envista’s analysts flag the visual errors. Fonts clash. Text sizes vary. Spacing looks odd. Timestamps defy logic, and carrier or battery icons fail to match the device. Blurry patches around text point to paste work, and odd edges betray pasted elements. Context tells follow the same logic. A post that no official account or newsroom record matches stays unconfirmed. The strongest check lives on the platform, not inside the picture. The firm’s guide lists the full set of tells for text fakes.
What happened when an embassy shared a fake?
The record supplies a clean case. In August 2022 a fake screenshot carried the name of Nobuo Kishi. He was then a special adviser to Japan’s prime minister. The image blamed Ukraine for strikes at the Russian-held Zaporizhzhia nuclear plant. The Russian Embassy in the United Kingdom posted it as if real. Kishi then confirmed he had never written it. The image registers false. The denial is confirmed. The origin of the fake was never publicly named. This connects to our earlier piece, No, the Pope never endorsed Trump — the claim's actual origin.
How do you check a screenshot before sharing it?
Verification desks run a short sequence. Every step is open to any reader.
- Find the post on the account itself, in the app or on the site.
- For older dates, use the platform’s advanced search for that account.
- Check the Internet Archive, which stores regular captures of major accounts.
- Run a reverse image search to find earlier fact-checks of the same image.
Engines such as TinEye and Yandex join Google in that search. The Global Investigative Journalism Network guide walks through each step with worked examples. Its test fake under a major broadcaster’s name died in a timeline search. Seconds of checking beat a confident share.
What do platforms do about fabricated media?
X’s authenticity rules ban misleading media. The policy defines manipulated media as material passed off as authentic while significantly and deceptively altered or fabricated. Enforcement spans labels, reduced reach, and removal. The policy names its own limit too. Where the platform cannot decide if media mislead, it may take no action. That gap keeps manual checks necessary.
Conclusion: treat every screenshot as a claim
A screenshot proves that someone produced an image. Nothing more. The Kishi case shows how far a flat image travels when a real institution forgets that. The checks take minutes. Find the live post. Search the archive. Reverse-search the picture. If the original never turns up, the claim stays unconfirmed. An unconfirmed claim does not need a share. Readers following this should also see Trace a viral photo's origin with reverse image search.
