Long before AI, photos already lied to us

An American postcard appearing to show the largest ear of corn ever grown. Image created by photographer W.H. Martin in 1908. (Courtesy Rijksmuseum via CNN Newsource)

A picture is often said to speak a thousand words, but do we still trust it to tell the truth?

The internet, editing tools, social media and — of course — AI, have made us increasingly aware that when it comes to photography, looks can be deceiving. Fabricated images, like that of the late Pope Francis sporting a snow-white puffer coat or U.S. President Donald Trump’s purported police mugshot, often go viral after capturing the public’s imagination.

• Pope Francis the fashion icon? Detecting AI images reaches ‘uncanny valley,’ cybersecurity expert warns

But while the technology that lets us create pictures of breakdancing babies and gangster cats is constantly evolving, doctoring images is nothing new — as an upcoming exhibition at the Rijksmuseum in Amsterdam demonstrates.

Opening Friday, “Fake!” shows how visual illusions have been created since the mid-19th century.

“We all talk about AI nowadays,” the exhibition’s curator, Hans Rooseboom, told CNN in a video call. “We’re used to Photoshop and other digital ways of altering images, but we wanted to show that it’s always been the case, since the very early days of photography.

“People have always had the tendency to play around with all the possibilities photography offers, both with the camera and in the darkroom, or with scissors and glue in a non-digital way.”

The show features 52 images from the museum’s collection dating from 1860 to 1940, all of which were devised using collage or montage. To create a photocollage, the artist physically cuts and pastes images together. In a photomontage, multiple pictures are combined and then rephotographed.

Like much of what we see from AI today, many of these early images show obviously fantastical scenes — like a man pushing a giant version of his own head in a wheelbarrow, or an enormous ear of corn being dragged along by horse and cart.

But in an era when going viral was not yet a thing, why did early photographers go to such lengths to create false images?

“Why wouldn’t people fake photographs?” asked Rooseboom. Photography “has never been realistic,” he said, particularly in the 19th century, when people were “more accustomed to seeing paintings, prints, drawings that do not tell the literal truths.

“People were only slowly getting used to photography and maybe slowly getting used to the idea that photographs could be more realistic than other images.” But, he added: “There’s very little comments from the time, so we hardly know the audience’s reactions to what they saw.”

The overwhelming motive for the early fakes was to provide entertainment — about three-quarters of the images in the exhibition were created for this purpose, Rooseboom said. Others were created for advertising or to make a political statement.

John Heartfield, the pseudonym of German artist Helmut Herzfeld, was a leading photographic satirist who was fiercely opposed to Hitler and his Nazi party. Heartfield’s 1934 image, used on the front cover of the left-wing Workers’ Illustrated Magazine (Arbeiter-Illustrierte-Zeitung) shows Joseph Goebbels, chief Nazi propagandist, as Hitler’s barber.

“It’s Hitler, but Goebbels is turning him into (Karl) Marx in order to attract the workers’ electorate,” Rooseboom said.

“Heartfield is both the best-known and, I think, the smartest person to have used photography to mock Nazism and all that the regime did, and to try to warn people about all the dangers that were looming or already taking place.

“It’s very interesting because that kind of satire is still very prevalent, if not more so than ever.”

In contrast with the early flights of fantasy and satire, photojournalism really only began to evolve in the inter-war period, and with it came a new expectation on photography to be truthful.

“People were only starting to get used to seeing a lot of photographs in the 1920s and 1930s with popular magazines,” said Rooseboom.

“There was no mistrust [prior to that period] because people were only used to seeing hand-drawn images, so only slowly the idea crept in that photography could and should tell the truth.”

In about three-quarters of the images featured in the exhibition, the fakery is “really clear,” said Rooseboom — giving the example of someone who appears to have performed a theatrical decapitation — but in some, it’s harder to detect how the manipulation was done.

For example, Rooseboom pointed to a “little postcard from an aviation show somewhere in L.A., with a lot of airplanes in the air. But the audience is not paying attention, and they are very close together. That simply must be a montage because that cannot have taken place in reality that way.”

“I always wonder if people back then would have seen through the trick or not,” Rooseboom said.

“We see more photographs every single day than most people in the 19th century would have seen in their whole life. So we are more or less used to looking at and judging photographs. Maybe it was much harder to distinguish between [what was] real and not real.”

In many instances, the images were created by anonymous photographers and reproduced as postcards. The postcard of a man wheeling his own head was made using an “amateur trick” of photomontage, according to Rooseboom. This would have involved combining several negatives, either by printing them together in a darkroom or cutting and pasting and then rephotographing them.

“It was described in various magazines and little booklets from the 1890s both in France and elsewhere. So you could learn this trick by following a kind of recipe,” he said of the image, thought to have been produced between 1900 and 1910 by an anonymous artist.

Peter Ainsworth is a course leader for the MA in Photography and Digital Practice at London College of Communication. He told CNN that artists who digitally manipulate images today often do so to make a point. “It’s often used as a satire,” he said, adding that creators seek to provide “a critical voice towards the problems inherent in the technology.”

The artist’s motive should also influence how we judge their work, he said, adding: “It’s to do with how it enters a wider ecosystem.” To illustrate this, he gave the example of the AI-generated “Trump Gaza” video that emerged last year. The clip was created as satire by artist Solo Avital and his partner but made headlines when Trump himself posted it online.

“So you have an artist who’s being critical of a particular position being utilized by the position that they’re criticizing,” said Ainsworth.

Elsewhere, the artist behind the popular Hey Reilly Instagram account, which pokes fun at celebrities with AI-altered images, told CNN they started out wanting “to make myself and my friends laugh.”

“Over time, I became more interested in what the work was reflecting back at us: our obsessions with status, celebrity, consumerism, and the way brands and faces function almost like a visual shorthand now,” said the artist, who asked only to be identified as Reilly.

“The stuff I make for Instagram is really for an in-the-know audience, a kind of digital fan club who get the joke straight away. If someone thinks the fakery is about trying to fool people, they’ve completely missed the point.

“We still have this deep-rooted sense that ‘the camera never lies’ — you can see that in how worried people are about AI images, especially in politics. Fakery only works because our eyes and brains are still wired to trust photographs.”

The “debate around dishonesty in AI and fakery” is “aiming at the wrong thing,” they said. “The fake images exist to point people back to the medium. It’s the power and influence of digital platforms, and the motivations of the people who own them, that we should probably be paying closer attention to.”

By Lianne Kolirin, CNN