When Apple announced the new iPhone 18 Pro and Pro Max, they also introduced something that, as a photographer, I'm definitely interested in. It's called Apple Reference Image, and the idea is to give us stronger evidence about where a photograph came from, as in whether it's real or not.
I can see some real benefits to that, particularly when somebody looks at a photograph and immediately dismisses it as AI, but there's one detail that raises a few questions; it's optional. You have to choose to use it.
So what happens when somebody doesn't? Have they got something to hide, did they forget to switch it on, or were they simply using a camera that doesn't offer it? That's the part I think we need to talk about alongside the technology itself, because how we interpret this could be just as important as what it does.
What Apple has actually announced
Apple says the iPhone 18 Pro's main camera captures signed sensor data in reference mode, and its Private Cloud Compute system then develops that into an unaltered reference image. You can view that reference alongside the main photograph in Photos and compare the two, and Apple has also announced support for developers to enable reference viewing in other apps.
There are launch restrictions too. Reference capture won't initially be available in the EU, and it won't be available at all in China.
The useful idea here is having something to go back to when you want to examine a finished picture, but a comparison isn't necessarily a complete record of every editing step. I wouldn't assume we'll get a list telling us which tool was used, in which application and in what order.
Where it could help
Imagine you've made a photograph of something unusual. Perhaps the light was extraordinary, or you happened to catch a moment that looks almost too good to be true. You share it and somebody says straight away that it's AI, and at that point "no it isn't, I was there" might not get you very far.
A trustworthy reference could give you something more useful to put in front of them, but it won't make everybody believe you. Some people have already made their minds up, whatever evidence you show them; for somebody who actually wants to examine the picture, though, additional evidence could help.
I can also see a potential benefit for picture editors and clients. If the question is whether something significant has been added or removed, being able to compare the finished picture with a trusted reference could make that conversation a lot clearer.
Editing isn't the same as dishonesty
This is the part that interests me most, because I spend a lot of my time teaching Photoshop and Lightroom. Editing a photograph doesn't, by itself, tell you whether somebody is being dishonest.
Take a portrait. I might adjust the white balance, lift the shadows, remove a temporary blemish, or go quite a bit further than that. Those are changes, of course they are, but what matters is what the photograph is for and what the changes actually do and what I'm telling you about it.
If I remove a person from a news photograph, that could change your understanding of what happened entirely. If I make a composite and tell you it's a composite, you're being given the information you need to understand the work. That distinction matters.
So I'd like to see tools like this encourage more useful conversations about editing. What changed, did it alter the meaning, and was the picture presented honestly? Those questions tell us far more than simply asking whether Photoshop was involved.
The trouble with voluntary
My first concern is that the people who want to show how their pictures were made have a reason to take part, whilst somebody who wants to conceal what they've done can simply choose not to. That doesn't make the feature pointless, but it does limit what we can conclude from it. We could end up with some pictures that have supporting evidence and many more where we still don't know, and the danger is how quickly "we don't know" turns into "it must be fake".
Imagine somebody captures a really important moment but hasn't enabled the feature. Perhaps they didn't know it existed, or didn't expect to photograph anything important that day; they saw something happening, picked up their phone and took a picture, which is hardly unusual behaviour. Should we trust their photograph less because they didn't think about authentication beforehand? I'd be very uncomfortable with that.
Then there are people using older phones, different cameras or equipment that simply doesn't support this system. Their photographs don't become less honest because somebody else bought a newer phone, and I wouldn't want us drifting towards a situation where you're expected to own certain kit just to have your work taken seriously.
The obvious response is to have it switched on for everybody, and that's a fair point, but I also understand wanting a choice. People use their cameras for all sorts of things, family pictures and visual notes and creative experiments and work in progress, and they may not want an extra reference workflow attached to everything they shoot.
Before deciding whether it ought to be on by default, I'd want to understand the practical experience. What happens when you've got no connection, how much extra storage is involved, and how long does the process take? Those are questions worth testing, and I'm not suggesting they're established problems. If something is easy to use and clearly useful, people have a reason to adopt it; if it adds confusion or friction, a lot of people won't bother.
What happens after you press the shutter
There's also the journey a photograph takes after it's been made. You might edit it, export it, send it in a message, upload it somewhere, or have somebody else download it and pass it on. What happens to the supporting evidence along the way? Can the person at the end find it, can they view it with the software they already have, and will they understand what they're looking at?
A feature can work perfectly for the photographer and still be of limited value to the viewer if the evidence is hard to get at. That's something I'd want to test properly, taking a photograph through a normal editing and sharing workflow and then seeing what the recipient can actually check.
I'd also want to see how clearly the interface separates missing evidence, a failed check and a successful check, because those are completely different situations. If they all get read as a simple real or fake verdict, we're going to create a lot of confusion.
A true photograph can still mislead
There's one final limitation that no amount of enthusiasm about authentication should make us forget; a camera can faithfully photograph a misleading situation. I can stage a scene, choose a viewpoint that leaves out something important, or put an inaccurate caption underneath.
Imagine I photograph a small group of people waiting outside a building. One caption says they're waiting for it to open, another says they're protesting against its closure. Same picture, completely different stories, and checking where the image came from doesn't by itself tell you which caption is accurate.
That's a recognised limitation of photo provenance more broadly; evidence about an image's history can't, on its own, establish whether the content is factual. We still need context, which means asking who published the picture, what they're claiming, and what other evidence supports it.
Where that leaves me
I'm definitely interested, and I can't wait to see how it works in practice. Giving photographers a stronger way to support their pictures could be really valuable, and so could making it easier to examine meaningful changes, but I think we need to be careful about what we conclude.
If a reference image is available, use it as evidence and understand what it does and doesn't establish. If it isn't, ask questions rather than jumping straight to an accusation.
For me, that's the biggest concern with the voluntary approach. The people who opt in may gain a useful way to support their work, whilst everybody else could find themselves under suspicion for no good reason.
