Ask Onix
How AI alters the photos we take
Modern smartphones use advanced algorithms to enhance images, sometimes filling in details that weren't originally captured. While these tools create sharper, more appealing photos, they also raise questions about authenticity and how we remember moments.
The Moon illusion: Samsung's AI trick
A viral test revealed Samsung's Galaxy phones can generate a crisp image of the Moon even when pointed at a blurry, pixelated version on a screen. The company's "100x Space Zoom" feature uses AI to recognize the Moon and artificially add craters and shadows. Samsung calls this a "detail-enhancing function," but critics argue it fabricates reality.
Computational photography: The invisible editor
Every smartphone photo undergoes extensive AI processing before it's saved. Ziv Attar, CEO of Glass Imaging and a former iPhone Portrait Mode developer, explains that phones don't just capture light-they guess what a better camera would see and reconstruct the image. "It's called computational photography," he says.
A Samsung spokesperson defended the practice, stating AI features aim to "enhance image quality while preserving authenticity" and can be disabled by users. Yet even with AI toggles off, phones still apply noise reduction, color correction, and High Dynamic Range (HDR) blending to improve photos.
When enhancement becomes distortion
Apple's Deep Fusion and similar tools use neural networks trained on millions of images to sharpen details, but critics say the results can look unnaturally smooth or plasticky. Some users report bizarre distortions, like AI-generated artifacts in fine textures, leading them to revert to older phones for more "realistic" photos.
"At Apple, our focus has always been to help users capture real moments so they can revisit memories as they experienced them."
Apple spokesperson
Lev Manovich, a digital culture professor at City University of New York, notes that automation makes professional-grade editing accessible to amateurs. However, phones now make creative decisions-like adjusting facial features or blending multiple shots-that users may not notice.
Cultural divides in AI beauty filters
Phones in Asian markets often apply aggressive AI "beauty filters" by default, smoothing skin, altering facial features, or even adding details like eyebrows or eyeballs. Attar describes these as "pure hallucination." In contrast, U.S. brands like Apple and Google have limited or disabled such filters, citing mental health concerns.
Manovich compares modern AI retouching to historical photo editing but calls generative AI's ability to add nonexistent details "radically new." Research suggests these edits can distort self-perception and plant false memories.
Moments that never happened
Google's "Best Take" feature lets users combine faces from multiple group photos into one "perfect" shot-a moment that never occurred. While Google frames this as fulfilling user desires, Rafał Mantiuk, a graphics professor at the University of Cambridge, acknowledges the ethical tension: "It's a group photo, not evidence for a crime."
Most AI enhancements can be disabled, but accessing truly raw, unprocessed images requires digging into "Pro Mode" settings or third-party apps like Adobe Lightroom. These unfiltered photos often appear noisy and soft, but Manovich argues they're worth exploring to understand what phones normally alter.
The future of photographic truth
As AI blurs the line between documentation and creation, the debate over authenticity intensifies. While smartphones democratize high-quality photography, their algorithms also reshape how we see-and remember-the world.
