From Pixels To Illusions
The Truth About AI-generated images
AI generated images have revolutionised technology, however, there have been several questions I’ve been asking myself when it comes to AI generated images or videos.
“How are they made?”
“How did it look so fake and is all of a sudden difficult to tell apart?!”
— How AI creates images
similarly to ChatGPT’s style of human-like responses, generative AIs create images by relying on machine learning models (MLM) trained on a variety of images.
“Convolutional neural networks (ConvNets) specialize in the ability to identify objects and patterns in data. The neurons are a specialized form that works in a similar manner as the human eye. Although not as complex as the human brain, the machine can recognize an image in a way similar to how humans see.”
These models don’t just “copy and paste” existing pictures.
Instead, they learn from millions of examples, replicating and picking up on shapes and colours on objects.
Once trained, the AI can generate completely new images by combining what it has learned in unique ways, almost like an artist painting from memory, but with a mathematical brain.
(Honestly, this is very similar to how we talk with AI, like how it responds so naturally. It’s trained on human’s tone and replicates it)
— The Scary Advancements
The more advanced the AI model is (like the new GPT-5), the more natural the images look.
That’s why we’ve gone from strange AI '“art”, to hyper-realistic photos and videos in just a few years.
The training datasets have grown massively, and the algorithms have become better at mimicking the natural position light, perspective, and even the subtle imperfections of real-life photography…..
— Thoughts
What do you think about this?
Will AI’s images continue to advance to the point where we can’t tell real from fake?! And will this lead to world disasters if given to the wrong hands?
Thanks for reading The Teen Index! Subscribe for free to receive new posts and support my work.
WORK CITED
“AI-Images.” CSB/SJU Library Guides, College of Saint Benedict and Saint John’s University, https://guides.csbsju.edu/AI-Images
There’s something rather wonderful about how they mimic our vision system, even if they’re not quite as nuanced as the human brain-yet still powerful enough to recognise images in their own way
I was reading recently that some ConvNet models now exceed human accuracy in specific pattern-matching tasks, especially when they’ve got enough labelled data to practise on (according to a blog post from DeepMind) —and that blew my mind
This was such an interesting breakdown. I like how you compared it to an artist painting from memory that’s exactly how it feels when I see AI images that look too real, almost like they’ve borrowed our collective memory. The part that rattles me most is what you ended on: the fact that we’re reaching a point where our eyes might not be able to tell truth from invention. It makes me wonder if the future of the internet will need some kind of watermarking system just so we know what’s authentic. Otherwise, like you said, in the wrong hands… yeah, it could spiral fast. 🌀