Games

People who do not exist created with Veo3 invade social networks

What makes these videos so popular, most likely, is their ability to depict everyday scenes with a visual realism that, until recently, seemed out of reach

29 May 2025

3' min read

There was a time, not too long ago, when AI-generated content was recognisable at a glance: too smooth faces, unnatural movements and blurred details. Today, however, we are facing a turning point that makes it impossible to distinguish the authentic from the constructed as the web is increasingly invaded by AI-generated content: images, voices, texts and now even videos that look real but are not.

Already a few months ago, Ghibli-style content, but created with artificial intelligence and which many had mistaken for new productions by the Japanese studio, had spread on social networks, while today Veo 3, the model recently launched by Google, is capable of generating videos so realistic that they completely, or almost completely, eliminate the boundary between the real and the fake.

From TikTok to Instagram via X and YouTube, over the last few weeks these contents have been populating the most widely used social platforms, collecting millions of views and user comments that highlight astonishment and concern as, in many cases, they even generate a 'disturbing' effect.

What makes them so popular is most likely their ability to depict everyday scenes with a visual realism that, until recently, seemed out of reach for a model. These are no longer abstract clips, but the faces are expressive, the movements natural, the lights and shadows handled with a precision that is unmatched by film products, and which is redefining the perception of digital content.

An example is the video made by László Gaál showing people busy shopping at a car show: 'Before you ask: yes, it's all artificial intelligence here. The video and audio are both generated from a single text prompt using GoogleDeepMind's Veo3. Whoever is 'cooking' this model, let them cook!" the creator writes in the caption of the video posted on X:

Before you ask: yes, everything is AI here. The video and sound both coming from a single text prompt using #Veo3 by @GoogleDeepMind .Whoever is cooking the model, let him cook! Congrats @Totemko and the team for the Google I/O live stream and the new Veo site! pic.twitter.com/sxZuvFU49s
— László Gaál (@laszlogaal_) May 21, 2025

And then there is the video posted by @HonestBlogging on X of the lecturer busy explaining Generation Z slang to a class of 'baby boomers' who show interest in listening and taking notes:

Veo 3, il generatore AI di Google diventa uno strumento per realizzare video virali

With over 25 million views, the clip of Will Smith eating spaghetti made by @javilopez and posted on X also went viral:

Just got access to Veo 3 and the first thing I did was try the Will Smith spaghetti test. SOUND ON pic.twitter.com/y0CiZwNxgM
— Javi Lopez ⛩️ (@javilopen) May 22, 2025

https://x.com/javilopen/status/1925495026903380358

While, a genre that is very much in vogue is that of street interviews or people who state with conviction that they are real and not the product of artificial intelligence: Hashem Al-Ghaili, one of the best-known creators, has published a series of clips in which the protagonists not only deny that they are produced by an algorithm but also come to reflect on their own identity.

Then there are those who, like Matt Shumer, thought of doing the unthinkable: a classic video of a youtuber showing and commenting on Fortnite gameplay, registering over 2.5 million views:

Veo 3, i video generati dall’AI diventano virali sui social

And if evolution on the one hand has ignited the enthusiasm of creatives, marketers and influencers who see artificial intelligence as a useful tool to speed up their work and cut costs, on the other hand, the enthusiasm coexists with a sense of disquiet: should the generated videos become indistinguishable from reality, what consequences await us in terms of misinformation and trust in visual content? Some experts, such as those of the AI Now Institute, have already sounded the alarm in a report published in 2023 arguing that without clear regulations, these technologies could be used to create false testimonies or misleading content