Apple wants to prove your iPhone photo is real with a new hardware-level authentication feature

Key points

• Apple will launch a feature called Reference Image with the iPhone 18 Pro and Pro Max later this month.
• The camera signs raw sensor data at the hardware level, then Apple's Private Cloud Compute turns it into an unalterable record stored in Photos.
• Developers can apply for access to a Reference Image API across iOS, iPadOS and macOS to build verification into third-party apps.
• EU users won't get Reference Image at launch, though they'll be able to view and develop reference images once iOS 27, iPadOS 27 and macOS 27 arrive.

What is Apple's Reference Image feature, exactly?

It's a way to prove a photo came straight from a camera, untouched by AI. Switch the iPhone 18 Pro into Reference mode and the sensor signs the raw data it captures, attaching a unique digital fingerprint to every pixel before any editing can happen.

Apple then sends that signed data to Private Cloud Compute, a system where processing happens on Apple's servers without Apple itself being able to read the content. The result is a locked reference image in your Photos app. Compare it later against any edited version and you'll see exactly what changed.

The company disclosed the feature in detail to The Verge AI.

Why does this matter right now?

Because telling a real photo from a generated one is getting harder. AI image tools now produce scenes indistinguishable from genuine photographs, and that has obvious consequences for news, courts and everyday trust online.

We've been tracking this problem since August: our story on Apple's earliest Reference Image prototype on 11 August 2026 found the feature in an iOS 27 test build, and a separate piece that same week covered Google's parallel move on AI watermarks.

Several industry standards already try to tackle this. SynthID, developed by Google DeepMind, embeds invisible watermarks into AI-generated content. C2PA, the Coalition for Content Provenance and Authenticity, attaches a tamper-evident record of how an image was made. Meta uses a similar approach called Content Seal. All of them tag images after the fact, or label AI-made content as such.

Apple's approach is different. Rather than labelling AI output, it certifies camera output at the moment the shutter fires. Authentication baked into hardware is harder to fake than a tag applied later in software. Whether courts or news organisations actually build their workflows around it is the question worth watching.

Who can use it, and when?

The feature arrives with the iPhone 18 Pro and Pro Max later this month. It requires the new camera sensor those phones carry, so older iPhones won't get it.

Developers building photo or journalism apps can apply for Reference Image API access, letting their apps display and verify signed photos. That opens the door for anyone with a professional reason to confirm an image's origin.

One catch: buyers in the European Union won't get Reference Image at launch. Apple hasn't explained why, though the EU's Digital Markets Act has previously complicated Apple feature rollouts in the region.

For everyone else, the upshot is straightforward. Take a photo in Reference mode and you'll have something genuinely new: a credible, hardware-backed record that the image hasn't been touched.