Asger Risom
Phones have always shaped how moments are captured.
But as artificial intelligence takes on a larger role in photography, it is quietly reshaping something deeper: how people remember the world itself.
Every time you take a photo with your phone, artificial intelligence is deciding what that moment should look like. The process is largely invisible, happens in milliseconds, and increasingly goes far beyond correcting lighting or blur.
Modern smartphones use layers of algorithms and AI models to reconstruct scenes, filling in missing details, smoothing faces, adjusting colours and, in some cases, inventing visual information that was never there. The result can be beautiful. It can also subtly alter how reality is recorded and recalled.
When photos stop documenting
One of the most striking examples comes from Samsung’s Galaxy phones and their “100x Space Zoom” feature. The devices can produce detailed images of the Moon that far exceed what their tiny lenses can physically resolve.
In a widely shared test, a user pointed a Samsung phone at a deliberately blurry, pixelated image of the Moon displayed on a computer screen. The phone generated a crisp lunar photo complete with craters and shadows that did not exist in the source image.
Samsung describes this as a “detail enhancing function.” Experts say it works because the AI recognises the Moon and fills in expected features based on training data, not direct observation.
What your phone actually captures
According to Ziv Attar, chief executive of Glass Imaging, phones rarely capture a single image when you press the shutter.
“When you click capture on a phone, you don’t just capture one image, you actually take anything between four to 10 images usually,” he told the BBC. Those images are blended together using techniques such as noise reduction, colour correction and high dynamic range processing.
Apple’s Deep Fusion system goes further, using neural networks trained on millions of photos to identify objects and process different parts of an image separately. The goal is a clearer, more pleasing result than a raw snapshot.
Aesthetic decisions by default
Critics argue that these systems often overshoot. Photos can take on a plasticky or watercolour-like appearance, and zooming in can reveal distortions that resemble AI hallucinations.
Some phone models, particularly those sold in parts of Asia, apply aggressive beauty filters by default. “It’s pure hallucination,” Attar said, describing features that redraw eyebrows, smooth skin or add facial detail that the camera never captured.
Rafał Mantiuk, a professor of graphics and displays at the University of Cambridge, said every phone now has its own visual “style.” “Pixel phones have a style. Apple phones have a style. It’s almost like different photographers,” he said.
Memory and perception
The issue goes beyond aesthetics. Researchers have found that AI-edited images can influence memory and self-perception, planting false recollections or altering how people feel about their appearance.
Google’s Pixel phones, for example, offer a “Best Take” feature that merges faces from multiple group photos so everyone looks their best. The resulting image may feel truer to how users want to remember the moment, but it depicts a scene that never actually occurred.
As Mantiuk put it, that may be acceptable for a group photo, even if it would be unacceptable as evidence.
Seeing what the camera saw
Users can disable some features like HDR or beauty filters, but truly unprocessed photos often require special modes or third-party apps. Raw images are noisier, flatter and less flattering, but they show what the camera sensor actually recorded.
Lev Manovich, a professor of digital culture at City University of New York, argues that occasionally shooting raw images can be valuable. It helps users understand what their phones normally change and what their photos really represent.
In an era where AI quietly edits not just images but experiences, the question is no longer whether photos are “real,” but how much invisible interpretation we are willing to accept in our memories.
Sources: BBC News
