Apple is ‘concerned’ about AI turning real photos into ‘fantasy’
Illustration by Haein Jeong / The Verge
The upcoming release of Apple Intelligence has spurred the iPhone maker’s own “what is a photo?” moment. In an interview with The Wall Street Journal, Apple software chief Craig Federighi said the company is aiming to provide AI-powered image editing tools that preserve photo authenticity.
“Our products, our phones, are used a lot,” said Federighi. “It’s important to us that we help purvey accurate information, not fantasy.”
iOS 18.1 brings a new “Clean Up” feature to the Photos app that can quickly remove objects and people from images — a capability that Federighi and WSJ reporter Joanna Stern noted is far tamer than editing tools offered by rivals like Google and Samsung, which can add entire AI-generated assets to images. Despite Clean Up’s limited capabilities, Federighi said there had been “a lot of debates internally” about adding it.
“Do we want to make it easy to remove that water bottle, or that mic? Because that water bottle was there when you took the photo,” Federighi said following a demonstration of Clean Up being used to remove items from the background of an image. “The demand for people to want to clean up what seem like extraneous details to the photo that don’t fundamentally change the meaning of what happened has been very very high, so we’ve been willing to take that small step.”
Federighi said that Apple is “concerned” about AI’s impact on how “people view photographic content as something they can rely on as indicative of reality.” It’s a subject we’ve spoken about frequently here at The Verge. Editing tools like Google’s Reimagine feature make it incredibly easy for a large number of users to add lions, bombs, and even drug paraphernalia to pictures using nothing but a text description, which could further erode the trust that people place in photography. Generative AI editing apps, when used nefariously, are making it easier to mislead or deceive others with increasingly convincing fakes.
Apple Intelligence (at least for now) doesn’t allow users to add AI-generated manipulations to images like competing services do. Any images that have been edited using the new object removal feature will also be tagged as “Modified with Clean Up” in the Photos app and embedded with metadata to flag that they have been altered.
Apple isn’t alone in taking such precautions — the Adobe-driven Content Authenticity Initiative has a similar “Content Credentials” metadata system, for example, that aims to help people distinguish between unaltered images and AI fakery. That requires tech, camera, and media companies to voluntarily back it, but support is steadily increasing. It’s unclear if Apple’s own metadata system will support Content Credentials.
Illustration by Haein Jeong / The Verge
The upcoming release of Apple Intelligence has spurred the iPhone maker’s own “what is a photo?” moment. In an interview with The Wall Street Journal, Apple software chief Craig Federighi said the company is aiming to provide AI-powered image editing tools that preserve photo authenticity.
“Our products, our phones, are used a lot,” said Federighi. “It’s important to us that we help purvey accurate information, not fantasy.”
iOS 18.1 brings a new “Clean Up” feature to the Photos app that can quickly remove objects and people from images — a capability that Federighi and WSJ reporter Joanna Stern noted is far tamer than editing tools offered by rivals like Google and Samsung, which can add entire AI-generated assets to images. Despite Clean Up’s limited capabilities, Federighi said there had been “a lot of debates internally” about adding it.
“Do we want to make it easy to remove that water bottle, or that mic? Because that water bottle was there when you took the photo,” Federighi said following a demonstration of Clean Up being used to remove items from the background of an image. “The demand for people to want to clean up what seem like extraneous details to the photo that don’t fundamentally change the meaning of what happened has been very very high, so we’ve been willing to take that small step.”
Federighi said that Apple is “concerned” about AI’s impact on how “people view photographic content as something they can rely on as indicative of reality.” It’s a subject we’ve spoken about frequently here at The Verge. Editing tools like Google’s Reimagine feature make it incredibly easy for a large number of users to add lions, bombs, and even drug paraphernalia to pictures using nothing but a text description, which could further erode the trust that people place in photography. Generative AI editing apps, when used nefariously, are making it easier to mislead or deceive others with increasingly convincing fakes.
Apple Intelligence (at least for now) doesn’t allow users to add AI-generated manipulations to images like competing services do. Any images that have been edited using the new object removal feature will also be tagged as “Modified with Clean Up” in the Photos app and embedded with metadata to flag that they have been altered.
Apple isn’t alone in taking such precautions — the Adobe-driven Content Authenticity Initiative has a similar “Content Credentials” metadata system, for example, that aims to help people distinguish between unaltered images and AI fakery. That requires tech, camera, and media companies to voluntarily back it, but support is steadily increasing. It’s unclear if Apple’s own metadata system will support Content Credentials.