X Close

AI fakes herald the end of war photography

A Palestinian child in the aftermath of an Israeli bombing. Credit: Getty

October 31, 2023 - 11:20am

Are we entering a new age of CGI war propaganda? At the weekend Chei Wenhua, a reporter for the CCP-run China Daily propaganda outlet, tweeted an image purportedly from Gaza alongside the caption, “Atrocities endorsed by the collective West.” The image depicts a man helping five small children through blasted rubble: three on his back, a baby in his arms, another holding his hand.

Twitter’s Community Notes fact-checking group swiftly flagged the image as AI-generated, pointing out a number of suspect features — including an odd distribution of limbs, the wrong number of toes, and other “uncanny valley” quirks typical of AI imagery. Even so, the picture has received 1.1m views at the time of writing, and a glance down the quote tweets suggests at least some have taken it at face value.  

We should expect such propaganda imagery to become both more frequent and more difficult to detect, as AI imaging gets more sophisticated. And one disturbing potential consequence may be a critical loss of faith in the kind of reporting able to galvanise humanitarian efforts during wartime. 

The twentieth century was the great age of war reporting — and especially of its pity and horror. From Holocaust footage through Vietnam war reporting to the broadcasting of both Iraq wars, journalism stripped away older narratives about heroism, focusing on the suffering of innocents. This helped energise a wave of efforts to impose restrictions on warfare with the aim of minimising such casualties.

More recently, too, it has impelled a new kind of propaganda warfare, as belligerents compete for the sympathy of an implied global digital audience. This has been powerfully in evidence from the beginning of Vladimir Putin’s invasion of Ukraine, with reports and counter-reports of gains, losses, and atrocities eagerly circulated online by and for warring political tribes, well beyond the physical battlefield in a competition for international public sympathy. 

If secured, this public sentiment can tilt the scales on real-world decisions, for example in helping to grant political legitimacy for military or financial support from foreign allies. So it’s well worth the effort to try and shape it. And in the digital-era media battles that do so, a picture is proverbially worth a thousand words. But this twentieth-century relation between media and warfare has grown more ambivalent, especially as the world has become increasingly multipolar and the internet has de-centralised broadcasting. 

Today, a multiplying number of actors with an interest in shaping public opinion struggle for influence, across multiplying media outlets. The heavily youth-focused (and CCP-owned) platform TikTok, for example, has been linked to overwhelming support for Palestine among US youth. Russian TV produces fake BBC bulletins linking Ukraine with Hamas. And against this backdrop of intensified propaganda war, we must now also reckon with a rising tide of increasingly convincing AI imagery. 

The aggregate effect is already discernible in the Israel-Palestine conflict. As reports and counter-reports of atrocities circulate, it’s increasingly common for overseas supporters of the warring sides to each dismiss the other’s claims simply as fake: not just exaggerations but pure fabrication. 

Beyond the immediate conflict between Israel and Palestine, we should expect this new landscape to degrade the media’s relationship to humanitarianism. For the power of such calls on public empathy was always predicated on a shared belief that there was at least some relationship between the images in circulation and what’s happening in the real world. What happens, though, when fakery is so good we can no longer take this for granted? 

The end of trustworthy battlefield imagery is the end of wartime humanitarianism. If no one is willing to believe anything they see or read about foreign atrocities, because no one is willing to accept media evidence, we will slowly lose our grip on every postwar convention governing human rights during wartime. It is a sobering prospect.


Mary Harrington is a contributing editor at UnHerd.

moveincircles

Join the discussion


Join like minded readers that support our journalism by becoming a paid subscriber


To join the discussion in the comments, become a paid subscriber.

Join like minded readers that support our journalism, read unlimited articles and enjoy other subscriber-only benefits.

Subscribe
Subscribe
Notify of
guest

19 Comments
Most Voted
Newest Oldest
Inline Feedbacks
View all comments
Right-Wing Hippie
Right-Wing Hippie
8 months ago

I saw an ad for the new Google phone, in which a young woman took a picture of herself and her friends; the ad then showed her altering the photo using built-in software on the phone so that everyone was looking in the same direction. This was touted as a desirable feature. So the very idea that a photo in some way reflects objective reality is no longer seen as a benefit but rather a defect, to be corrected with technology. Why live in the real world when it’s so easy to correct the real world and make it more congenial?

Warren Trees
Warren Trees
8 months ago

I said the same exact thing when I first saw that TV ad. Why bother visiting a beach, see a sunset or enjoy a wonderful meal anymore when you can simply create a photo of anything and post it online, so everyone thinks you are a world traveler with unlimited resources. Crazy times.

Peter D
Peter D
8 months ago
Reply to  Warren Trees

Why go through the hassle and discomfort of traveling when technology can bring you anywhere in an instant. Stay home, be comfortable. Bread and games!

Laura Creighton
Laura Creighton
8 months ago

Maybe it is time to create the Institution of “Fair Witness” from Heinlein’s _Stranger in a Strange Land_ ?

Brian Villanueva
Brian Villanueva
8 months ago

In many ways we are re-entering the pre-photography age, an age when the veracity of an event you saw described could only be determined by the veracity of the person who told you.

It will take a lot of adjustment to handle this, but it will be a good thing. For too long, we have treated video and pictures as if they told the whole story. Now we will need to find people who actually know the whole story and can give it to us honestly, so that when we tell it to others, they will trust our account because we have never steered them wrong before.

In other words, anonymous-Twitter-dude is going to fail. Reputation once again matters.

John McWade
John McWade
8 months ago

Oooh, intriguing insight, thank you. The pre-photo world.

starkbreath
starkbreath
8 months ago

There’s a huge fallacy underlining your scenario: too many people are not interested in the truth, they just want a story that confirms to their own narrative, which must remain unchallenged and unchanged. The advent of AI imagery, still in the early stages, will only make this exponentially worse. It’s like arguments over Facebook, Twitter, Til Tok, et al: the only solution would be for people to stop using them, which is not going to happen.

Allison Barrows
Allison Barrows
8 months ago

Their AI skills are as sophisticated as their videos of shrouded moving “corpses” so commonplace with the Jihadi crowd. Subtle, they’re not.

Kate Madrid
Kate Madrid
8 months ago

Humanitarian concerns didn’t originate with photography. And it’s a net positive for everyone to start taking seriously the difference between an image and the world.

Steve Murray
Steve Murray
8 months ago
Reply to  Kate Madrid

Absolutely right. Photographic images have always been dependent on the time and the subjectivity of the photographer. The resulting “snapshot” may seem to convey something truthful, but the trillions of other images that could’ve been taken either before or afterwards might have revealed something equally truthful but different.

Gordon Black
Gordon Black
8 months ago

Maybe we have to go back to celluloid: faking original negative film is difficult.

Stuart Bennett
Stuart Bennett
8 months ago

We haven’t been able to trust images for quite a while now. The more there is to gain from an image the less it can be trusted.

Steve Jolly
Steve Jolly
8 months ago

I’ve been assuming everything on the Internet is fake since about 2005 when I became aware of a primitive non-AI tool called ‘photoshop’. Even before that, I was the weird kid that wouldn’t raise his hand in class when asked if ‘the US is the greatest country in the world’. I don’t buy the propaganda. I never drank the Kool-Aid. Part of it is my nature and part of it was being raised by two intelligent parents who did their thinking for themselves rather than have it be spoon fed to them by the government or the media. Nice to see that fear of the AI apocalypse has inspired others to the same level of cynicism that I achieved decades ago. If AI causes people to not trust what they see in the media and on the internet, that’s a point in its favor, because the manipulation of information to influence public opinion is old as dirt and never required AI any of the other times its been done throughout history. If nobody trusts the media, there’s a pretty hard limit on how much damage they can do. On the other hand, if everyone trusts them, they can exercise incredible power. I prefer the former, thank you very much. The proliferation and democratization of propaganda is, sadly, probably a net gain for society, given that such power was once reserved for governments and wealthy media magnates (read up on William Randolph Hearst sometime). If everyone is crying wolf, and half the people are saying it’s to the west, and the other half to the east, we know somebody or everybody is lying, and we know there may be a wolf somewhere, which on balance is about the same as if nobody said anything. This may seem depressing to the far too idealistic people of today, but this cynical libertarian finds it quite acceptable, and mildly humorous.

Jon Hawksley
Jon Hawksley
8 months ago

Step one – AI should lead to everyone questioning the veracity of all images and words online.
Step two – the abuse of AI will be driven by the desire to earn money by clicks which is funded by advertising.
Step three – a desire to find trustworthy souces of information will encourage people to follow advertisement free subscription services.
Step four – as subscribers find some sources are less trustworthy because content is tailored to maximise subscriptions they will tend to support not for profit sites, maybe funded as a mutual association.
Optimistic? I hope not.

Haydn Pyatt
Haydn Pyatt
8 months ago

It’s nothing new. Reality has always been a subjective experience modelled by what we want to believe. A schizophrenics view on what’s real is as valid as anyone else’s.

Warren Trees
Warren Trees
8 months ago
Reply to  Haydn Pyatt

I don’t know about that. Personally witnessing the dismemberment of a platoon after a bomb lands on them is a reality that is a little different than reading about it in a magazine from the comfort of our sofa.

starkbreath
starkbreath
8 months ago
Reply to  Haydn Pyatt

Spare us the postmodern bullshit already.

Matt Sylvestre
Matt Sylvestre
8 months ago
Reply to  starkbreath

Indeed – postmodernism and bullshit are one in the same (though the later can at least fertilize the garden)…

Duane M
Duane M
8 months ago
Reply to  Haydn Pyatt

I presume that you are speaking from your own experience as a schizophrenic. How else could you know?