Realtime AI News
Apple's Reference Image uses sensor signatures to show an iPhone photo isn't AI slop
Apple introduced Apple Reference Image, a feature that uses signed sensor data captured by the iPhone's main camera and Apple's Private Cloud Compute to produce a verifiable reference version of a photo. Reference images will first be viewable in the Photos app, with developer APIs to follow, and Apple said it will support the SynthID standard.

Apple introduced a feature called Apple Reference Image at its September 9 event, saying it can prove whether a photo taken with an iPhone is authentic, including whether it was altered by AI.
The mechanism is hardware-based. When a photo is captured, the main camera sensor also records signed sensor data, which Apple's Private Cloud Compute service turns into an unalterable image that can be viewed in the Photos app. Apple likens the result to a digital negative.
Users can compare that reference image with other versions of the same photo to determine whether any changes have been made. Apple said the capability is vital for photojournalists and photographers.
Reference images will initially be viewable only in the Photos app, but Apple is making APIs available so developers can integrate the feature into their own apps. The company also said it will support the SynthID standard, which helps people tell whether images were created or altered with AI.
The feature sits inside a broader camera-focused refresh. Apple described the iPhone 18 Pro as its biggest camera improvement in years, and highlighted the variable aperture on the new 48-megapixel fusion main camera, which gives users more control over how much light reaches the lens and should be especially useful for night photography.
To build that aperture, Apple said it developed a system of six laser-cut blades, each thinner than a human hair and made from a specialized polymer composite, along with a new rotor mechanism and a custom-designed lens.
The feature gives Apple a hardware-level answer to a problem the whole industry is wrestling with: proving a photo came from a camera rather than a generative model.
What to watch next is how quickly developers adopt the APIs, how far SynthID support extends across Apple's devices and workflows, and whether sensor-signed provenance becomes a default expectation for newsroom-grade photography.
Why it matters
By tying provenance to camera hardware rather than software alone, Apple moves the authenticity question into the device itself, which could matter for journalism and for how platforms treat photos as evidence.
Nearby Updates
All09/10, 01:36
Apple's new CEO says the best AI device is still the iPhone
At Apple's "Surprise and Shine" event, new CEO John Ternus described the iPhone as the ideal "intelligent personal hub" and stressed that Apple Intelligence runs on device whenever it can. The pitch signals that Apple will compete in AI through hardware it already ships and a privacy-first, on-device approach rather than a new AI gadget.
09/10, 01:00
Paul Christiano joins the OpenAI Foundation Board
OpenAI said Paul Christiano is joining the OpenAI Foundation Board and its Safety and Security Committee. His experience in AI alignment, safety and standards will inform the foundation's governance work.
09/10, 03:21
Apple’s first foldable, the Duo, has a hinge built with AI and 3D printing
Apple unveiled the Duo, its first foldable phone, at its Wednesday event and said the device’s critical hinge was designed and built with AI algorithms and 3D printing. Hardware chief Johny Srouji said each hinge is laser-scanned and then built up with up to 25 3D-printed photopolymer micro layers to ensure perfect alignment.
09/10, 00:05
Superintelligence is coming. Should we let it?
Superintelligence is coming. Should we let it?. AI companies have been talking about superintelligent AI like it’s inevitable, but recent safety incidents like OpenAI’s Hugging Face breach are demonstrating the potential dangers of deploying AI systems that are more capable than humans. So what happens when...