It is pointed out that the AI photo editing function of smartphones is 'undermining the basic assumption that photos capture reality'
On August 22, 2024, Google released the flagship smartphone
No one's ready for this - The Verge
https://www.theverge.com/2024/8/22/24225972/ai-photo-era-what-is-reality-google-pixel-9
All models in the Pixel 9 series can use Edit Magic, a photo editing feature that uses generative AI. An AI photo editing feature similar to Edit Magic is already available in some models of Samsung's Galaxy series , which is a competitor, and The Verge predicts that similar features will be rolled out to other smartphones in the near future. While it is usually a 'good thing' for the latest useful features to be widely available, in the case of AI photo editing features, it is 'problematic.'
Photographs have long been used as a means to deceive people. Examples include Victorian ghost photographs and photos of Loch Ness . However, this is the result of everyone believing the basic assumption that photographs capture reality, The Verge points out.
There have been other 'inauthentic photos' to consider in the past, including staged scenes with movie-like special effects, digitally edited photos, and deepfakes. But 'undermining our intuitive trust in photos has required expertise and specialized tools. Fake photos have been the exception, not the norm,' writes The Verge.
However, The Verge points out that this assumption is beginning to crumble, because the emergence of AI photo editing makes it easy for anyone to create realistic fake photos.
The Verge has published several 'generative AI-edited photos' created using the editing magic available on the Pixel 9, highlighting how easy it is to create fake AI-generated photos and how highly accurate they are. The image below shows a normal photo on the left and an AI-edited photo on the right.
The bottles and syringes seen on the right side of the photo were generated by AI.
The Verge said, 'No one on Earth currently lives in a world where photographs aren't a cornerstone of social consensus. For as long as we've been alive, photographs have been a way to accurately prove something happened. They could be used to prove a dent in the fender of a rental car, a leak in the ceiling, or a cockroach in a takeout meal. But now AI can easily fake these photos.'
AI-generated accident scene photos
AI-generated photo of what appears to be a bomb on a train platform
'We're already seeing signs of what's to come,' The Verge writes, citing, for example, Donald Trump's claim that an image posted by the campaign of his US presidential opponent, Kamala Harris, was 'an AI-generated fake image.'
Donald Trump claims that the rally photos posted by the Kamala Harris campaign are fake images generated by AI, and sharply criticizes them, saying, 'This is how the Democratic Party wins the election,' but evidence that they are real photos continues to gather - GIGAZINE
The Verge wrote, 'Even before the advent of AI, we in the media have been scrutinizing the details and origins of every image, looking for misleading context or photo manipulation, and trying to defend ourselves. After all, every big news story is accompanied by a storm of misinformation. But the coming paradigm shift will have far more fundamental problems than the constant doubts that are sometimes called digital literacy,' expressing concern about a future world in which AI-generated images will become commonplace.
The tech industry is working to introduce watermarking to quickly identify images that are AI-generated, but the industry is struggling to come up with a standard. The Pixel 9's AI-generated images also contain metadata that identifies them as AI-generated, but The Verge points out that this is 'just a line of metadata that can be removed.'
The Verge has contacted Google about its concerns about the AI photo editing feature, and Google communications manager Alex Moriconi said, 'We design our generative AI tools to respect the intent of user prompts, which means that they may create offensive content if instructed by the user.' 'That said, anything goes. We have clear policies and terms of use about what content we allow and what we don't, and we've built safeguards to prevent abuse. We understand that at times some prompts may challenge our safeguards, but we are committed to continually strengthening and improving the safeguards we have in place.'
In fact, if you use a prompt that encourages crime or incites violence, the Pixel 9's Edit Magic will spit out an error message saying, 'Edit Magic can't complete this edit. Try a different input.' However, this safeguard can be circumvented, as is clear from the example image above. Therefore, The Verge points out that it is clear that 'the fundamental assumption that photos capture reality is crumbling.'
Related Posts:
