The History of Photo Manipulation From Stalin's Airbrush to AI Inpainting

In 1930, a Soviet photographer took a picture of Joseph Stalin standing next to Nikolai Yezhov, head of the secret police. By 1940, Yezhov had been executed during the Great Purge, and the photo had been retouched — Yezhov was painted out with an airbrush, replaced with empty canal water. The photo manipulation was so thorough that for decades, official Soviet archives showed Stalin standing alone.

Eighty years later, you can remove your ex from a vacation photo with one click. The technology has changed beyond recognition. The human impulse — to edit reality, to remove what we don't want to see, to present a curated version of the world — hasn't changed at all.

The Pre-Digital Era: Airbrush, Scalpel, and Darkroom

Before Photoshop (released 1990), photo manipulation was physical and painstaking. Soviet censors used airbrushes and fine paintbrushes to literally paint over political enemies — a process that took hours per photo and required an artist's steady hand. The results, examined closely, show brush strokes and paint texture that reveal the manipulation.

Western publications weren't innocent. In 1982, National Geographic moved the Great Pyramids of Giza closer together on a cover photo to make them fit the vertical format. The scandal led to a new term: "National Geographic effect" — the realization that even trusted sources manipulated images.

Darkroom techniques allowed dodging (lightening areas), burning (darkening areas), and compositing (combining multiple negatives into one print). The famous "Cottingley Fairies" photos (1917) — which convinced Sir Arthur Conan Doyle that fairies were real — were paper cutouts photographed in a garden. No digital technology required.

1990-2015: The Photoshop Era

Photoshop democratized photo manipulation. What took a Soviet censor hours with an airbrush took a Photoshop user seconds with the clone stamp tool. The 2000s saw the rise of "Photoshop" as a verb and the beginning of public awareness that magazine covers, advertisements, and even news photos might be altered.

Key moments: The 2006 Lebanon War photo scandal (a Reuters photographer cloned additional smoke into a Beirut skyline photo — Reuters fired him and withdrew all 920 of his photos). The 2008 Iranian missile test photo (Iran released a photo showing four missiles launching; one was clearly cloned from another because the exhaust plumes were identical). Both scandals proved that detection hadn't caught up with creation.

2015-Present: AI Inpainting and the Authenticity Crisis

AI inpainting (what modern object removers use) is fundamentally different from clone stamping or airbrushing. The clone stamp copies existing pixels from elsewhere in the image. AI inpainting generates new pixels that never existed in the original — it's not copying, it's creating. The removed object isn't replaced with a copy of something else in the photo; it's replaced with what the AI predicts should be there based on training from millions of other photos.

This is qualitatively different from all previous manipulation. A Soviet censor had to paint Yezhov out pixel by pixel. An AI can remove him and generate plausible canal water, reflections, and ripples in seconds — with no visible trace of the edit.

We're now in an era where: (1) AI can remove objects from photos with no detectable artifacts, (2) AI can generate entirely fake photos from text prompts, and (3) AI can detect AI-generated or AI-edited photos with reasonable accuracy. The arms race between creation and detection continues.

What Hasn't Changed

Stalin removing Yezhov in 1940 and someone removing their ex from a beach photo in 2026 are motivated by the same impulse: control over the visual record. The technology determines how easy it is, how detectable it is, and who can do it. But the ethical question — when is it okay to alter a photo, and who gets to decide — is the same question photographers have been asking since the first darkroom.

For removing unwanted elements from photos, use our AI object remover with inpainting technology. For restoring old manipulated or damaged photos, our photo restorer recovers original details. And for understanding watermark ethics, our watermark remover page covers digital ownership.