Real life is beginning to blend with fiction

One otherwise innocuous YouTube video has left many viewers questioning their reality, as a woman on a farm has used AI to blend reality with fiction with terrifying results.

While the generative AI revolution started out with ChatGPT and text-based responses, a significant part of the focus right now involves image and video generation with ever-improving results.

OpenAI took over social media with it's Studio Ghibli-style image conversions, and everyone from Google to Elon Musk's Grok are now offering users the ability to create short videos completely from written prompts.

One creator has used AI generations alongside real life footage to show quite how convincing the technology has become, and the result is equally fascinating and horrifying for its implications.

What does the video show?

A YouTube short posted by Photoshop and AI expert Madeline Salazar has drawn the attention of many, as her time on a farm is filled with numerous fake features as if we were watching someone in the matrix.

Madeline asks her viewers at the start of the video to guess whether prompted parts of each new scene are AI generated or real, and it's genuinely difficult and even nigh on impossible to accurately decipher between the two.

Most of the elements are indeed AI generated as a means to show off the impressive assimilation of the tech into real life footage, but she has thrown in a few legitimate objects into the scene to help prove her case even more.

One particularly scary one sits near the end of the video, where she asks people to figure out whether a bridge she's standing on is real. For this particular demonstration it is, but the trick was that the stream flowing underneath it isn't, as she jumps down to reveal a flat dirt field instead.

There are also moments where chickens moving freely around a field are revealed to be a complete fabrication, showing that something doesn't need to be static to seem real to our human eyes.

How have people responded to the video?

Countless comments have expressed their disbelief and fear in response to the video, with one user revealing that they "got every single one wrong," and another adding that "that was kind of alarming."

Others have even suggested that they were waiting for a twist when Madeline herself revealed that she'd been generated by AI for the entire video, and with the way things were going that wouldn't have been too out of the question.

Some can spot that it's AI in the small details, but it's hard to really notice on a first watch (YouTube/immadsal)

"Damn, if you look back at the things that were AI, you can tell if you really pay attention to small details," writes one user, but they point out that the bigger worry is how easy it is to ignore these details when viewing something for the first time with no prior knowledge of AI tampering.

"If you're just glancing and not expecting anything to be wrong, you'd probably never notice," they argue. "The real danger is going to be when it's indistinguishable."

The Trump administration has already pushed forward a key government bill that would specifically fight against instances of AI generated non consensual imagery, but there might need to be wider efforts to tackle the wave of AI images before it's too late to really know whether what you're looking at is real or fake.