iOS 27 Beta 5 includes references to an upcoming new feature to allow Apple to authenticate iPhone photos.
Not many iPhone users will need to use this for reasons we’ll get to in a moment, but the move is just the start …
AI-generated images and video have rapidly grown in sophistication and realism to the point where it can sometimes be hard to distinguish them from genuine photos. This fact is being exploited by both scammers and those intent on spreading disinformation.
An industry body has been set up to fight this, known as the Coalition for Content Provenance and Authenticity (C2PA). This provides a means of identifying genuine photos and enabling viewers to see how they have been modified, whether or not AI tools were used.
Over 500 companies have joined the C2PA initiative, while Apple for now appears to be going its own way with the Apple Reference Image (ARI) feature. Anyone who has taken a photo on an iPhone with ARI enabled can upload it to an Apple server for verification that it is genuine.
Assets found within the latest beta indicate that the ARI feature will be off by default, and that’s because most iPhone owners will never need to use it. It’s primarily intended for people like journalists and those involved in legal proceedings who may need to be able to produce independent proof that their photos are genuine.
But while ARI may have limited usage when it launches, features like this are going to become increasingly common as usage of AI image-generation grows. For example, the EU recently announced a code of practice on the transparency of AI-generated content intended to allow people to easily tell when content is not real. This applies to both images and text (see below).
This is currently a voluntary initiative, but one AI giant has already seen which way the winds are blowing and has announced support for it. Anthropic says that all future models of Claude will apply digital watermarking to both text and images. Notably, the company is doing this worldwide, not just within EU countries – a tacit acknowledgement that such requirements are likely to become mandatory globally in the not too distant future.
The most eye-opening requirement in the EU initiative is that AI companies must digitally watermark text output as well as images.
This can be done using invisible characters, such as the one I just used to replace the space between the words ‘invisible’ and ‘characters.’ This would survive copying and pasting, although would be very easy to defeat by using optical character recognition. For example, I have a utility on my Mac that allows me to do as easily as taking a screengrab (installed because of the highly annoying habit some companies have of tweeting statements as images).
Some are suggesting that Anthropic may go as far as using particular language patterns in Claude output in order to allow detection even for OCRed text. I could have very much to say about this, but that’s beyond the scope of this piece.
If this kind of thing does indeed become mandatory, then Apple will need to apply it to tools like Siri AI rewrites.
What are your thoughts on this trend? Please share in the comments.
FTC: We use income earning auto affiliate links. More.
Check out 9to5Mac on YouTube for more Apple news:
Breaking news from Cupertino. We’ll give you t…
Ben Lovejoy is a British technology writer and EU Editor for 9to5Mac. He’s known for his op-eds and diary pieces, exploring his experience of Apple products over time, for a more rounded review. He also writes fiction, with two technothriller novels, a couple of SF shorts and a rom-com!
