The War That Is Won First in AI — S. Rouzbeh

The War That Is Won First in AI

From Trump’s Strait of Hormuz to the Islamic Republic’s atomic bomb: when powers portray what has not yet happened

Sometimes the image of a war is made before the war itself.

In recent weeks, Donald Trump has published several AI-generated images and videos about Iran. On 31 August, he shared a video on Truth Social showing Kharg Island being destroyed. At the time, a US official told Reuters that no attack on Kharg had taken place. US Vice-President JD Vance also said that publishing the video had been «a message to Iran», not an announcement of a military operation.

A few days later, on 6 September, other images appeared on Trump’s account, including a map on which the Strait of Hormuz was marked with the words «NEW U.S. Territory». Other images also showed scenes of war, explosions and American military power. These posts are recorded in the American Presidency Project archive.

Almost simultaneously, an example of the same visual language appeared in the Islamic Republic’s media sphere. Saberin News published a video titled «Iran’s first hot nuclear test» and used the word «animation» in its description. The video was not an actual test of a nuclear weapon, but a visualisation of a scenario showing the hypothetical explosion of Iran’s first nuclear bomb.

Two sides of a conflict, two different narratives and one shared tool: creating an image of something that has not yet happened in the real world.

The question is: why?

When reality is not enough

Political propaganda and psychological warfare have a very long history. In wars, governments have used posters, radio, cinema, manipulated photographs and propaganda films to make their own power seem greater, their enemy weaker and victory closer.

AI, however, makes something else possible.

Today, there is no longer any need to win before showing a victory.

An island can be blown up before an attack; enemy ships can be sunk before they go under; political borders can be changed; or an explosion can be shown of a weapon that has not yet been used.

An image is no longer merely a record of what has happened; it can show what its creator wants to happen—or at least what they want others to take seriously as a possibility.

This is where the distance between «existing reality» and «desired reality» grows shorter.

Real power or the image of power?

It is tempting to call such images «an illusion of power» or a display of unattainable ambitions. But rather than explain much, this description makes a judgement about the psychological state of those who produced them.

The more precise question is something else:

Has AI enabled politicians and governments to create an image of power they do not yet possess in the real world?

In one image, the Strait of Hormuz becomes ‘America’s new territory’. In another, Iran has moved beyond debate and nuclear threats, and depicts the explosion of its first bomb.

In the real world, each of these scenarios would face immense military, political, legal and economic obstacles. In the artificial world, however, a few prompts are enough to create them.

Here, AI does not only invent an imaginary enemy; it also produces an idealised image of the ‘self’: I am powerful, I am in control, I can do this.

When people see a war before it happens

These images are not only about their creators. More important are the millions of people who see them.

Research in cognitive psychology has shown that images can affect human attention, understanding, memory and judgement. Even an image that is not, in itself, evidence for a claim can, in certain circumstances, make that claim more imaginable and acceptable to the mind.

There is a difference between reading the sentence ‘Kharg may be targeted’ and seeing an image of Kharg amid fire and explosions.

The first is a possibility.

The second gives that same possibility shape, colour, light, smoke and sound.

We might call this phenomenon—not as an established scientific term, but as a concept to explain this situation—‘pre-experiencing war’: a person experiences an image of an event that has not yet happened.

And for an image to have an impact, the audience does not always need to think it entirely real. Sometimes it is enough for them to ask themselves, for a few seconds:

‘What if this really happens?’

One image, several audiences

Such an image may serve several purposes at once.

For a domestic audience, it can be a display of power: ‘We are capable of doing this.’

For the other side, it can take the form of an unofficial threat: ‘This is the future that may await you.’

And for world public opinion, it can introduce a scenario into the news cycle. The media write about it, experts analyse it, and social media users share it again.

The Kharg example is interesting in this respect. Vance explicitly said that Trump’s video had been “a message to Iran”. In other words, the fabricated image could carry a real political message even without the operation taking place.

An interesting contradiction emerges here:

If everyone knows that an image is artificial, but everyone talks about the scenario it depicts, has the image failed?

Or has it achieved exactly what it set out to do?

War in people’s minds

Images of war do not affect political calculations alone; they can also influence public fear and anxiety.

Studies of real images of war, terrorism and violence have shown that widespread and repeated exposure to such images is associated with an increase in symptoms of psychological stress. In a large study following the Boston Marathon bombing, people who were exposed to six or more hours of media coverage of the incident each day showed, on average, more symptoms of acute stress. Later research has also emphasised the importance of the kinds of violent images people see.

But we should not draw conclusions from this research that it has not yet established. Most of these studies have concerned images of real violent events, not wars created with AI.

As a result, we cannot yet say with certainty exactly what repeated exposure to artificial wars will do to the collective psyche.

It may create fear.

It may normalise threats.

And perhaps, after frequent repetition, it may even lead to a kind of numbness towards images of war.

These are questions that still require further research.

The greater danger: when even our eyes are no longer evidence

Perhaps the greatest danger of political images created with AI is not that people will believe all of them.

A systematic review published in 2026, which examined 25 experimental studies on political deepfakes, suggests that such content is not necessarily more persuasive than other kinds of misinformation. Its more significant effect may be to create confusion and erode trust.

A study of 2,005 people in Britain reached a similar conclusion: when faced with deepfakes, people were more likely to feel uncertain than to be deceived outright; and this uncertainty was associated with reduced trust in social media news.

This may be the most dangerous stage.

A fake image is published: perhaps it is real.

A real image is published: perhaps it was made by AI.

The government denies it: perhaps it is lying.

The government confirms it: perhaps it is psychological warfare.

In this situation, it is not only truth that suffers; our confidence in our ability to recognise the truth suffers too.

Victory Before the War

In the past, war propaganda usually entered the fray after an event: a battle took place, and the propaganda apparatus tried to portray it as greater, more heroic or different from reality.

AI can reverse this order.

Now the image of victory can be created first.

The enemy has been defeated.

Its ships have been sunk.

Its island has been blown up.

Its borders have been changed.

Even the explosion of a bomb that has not yet gone off can be shown.

In this way, another arena has taken shape alongside land, sea and sky: a battlefield over people’s perceptions of the future.

Perhaps, in future wars, what matters will not only be who wins on the real battlefield; it may also matter who gets their image of victory into people’s minds first.

But three questions remain:

When leaders and governments can portray their desired future before it happens, where is the line between threat, propaganda and reality?

When people repeatedly see wars that have not yet happened, will they become more afraid, or gradually become numb to war?

And perhaps most importantly:

If one day an authentic image of a real war is published, how will people who have seen hundreds of artificial versions of it know that what they are seeing this time is real?