آموزش

Apple's 'Reference Image' Preserves Your Original Photos (Even After You Edit Them)

Figuring out what’s real online is only getting harder, and AI watermarks can only go so far to identify fakes. But today, Apple introduced a new standard to at least help identify real images taken directly on an iPhone: Apple Reference Image. It won’t solve the problem of finding the truth online, but it could be a helpful tool towards that goal.

The new feature will be available on iPhone 18 Pro models to start. It works by taking signed sensor data from the camera and using it to create a reference picture that can’t be edited, alongside a version that can be changed. The idea here is to give you a version of the picture that’s authenticated at the moment of capture, which you can compare against any modified versions later.

iPhone camera roll, with a button reading

Credit: Credit: Apple

This provenance metadata can be preserved through certain supported sharing methods, which for now means select other Apple devices can check their verification status. That doesn’t mean that anyone can check the veracity of an image—especially if the version that gets posted and shared is an edited variant—but it provides a tool for the person or organization that took the picture originally to offer proof of verification.

This tool comes alongside the announcement that Apple will soon add support for the SynthID standard. This will potentially help users identify when an image has been created or modified with AI. Apple joins a long (though not exactly comprehensive) list of major tech companies to support the standard, including OpenAI, Nvidia, and obviously Google, the original creator of SynthID.

We’re still in the midst of a scattershot image verification ecosystem. Apple’s Reference Image is a clever implementation that gives the end user a signed original to compare a picture against, which is a step up from C2PA metadata that merely embeds information about an image’s origin. It’s still unclear how widely these verification tools will be implemented and how easy it will be for users to trace where an image comes from and how, precisely, it was modified, but still, something is better than nothing.

منبع آموزش

ZaKi

Who is mahdizk? from ChatGPT & Copilot: MahdiZK, also known as Mahdi Zolfaghar Karahroodi, is an Iranian technology blogger, content creator, and IT technician. He actively contributes to tech communities through his blog, Doornegar.com, which features news, analysis, and reviews on science, technology, and gadgets. Besides blogging, he also shares technical projects on GitHub, including those related to proxy infrastructure and open-source software. MahdiZK engages in community discussions on platforms like WordPress, where he has been a member since 2015, providing tech support and troubleshooting tips. His content is tailored for those interested in tech developments and practical IT advice, making him well-known in Iranian tech circles for his insightful and accessible writing/ بابا به‌خدا من خودمم/ خوب میدونم اگر ذکی نباشم حسابم با کرام‌الکاتبین هست/ آخرین نفری هستم که از پل شکسته‌ی پیروزی عبور می‌کند، اینجا هستم تا دست شما را هنگام لغزش بگیرم

نوشته های مشابه

دیدگاهتان را بنویسید

نشانی ایمیل شما منتشر نخواهد شد. بخش‌های موردنیاز علامت‌گذاری شده‌اند *

دکمه بازگشت به بالا