Apple has rolled out support for SynthID watermarks in iOS 27.2, embedding a hidden marker into images generated through its Image Playground tool and many photos edited using Apple Intelligence. SynthID is a watermarking system developed by Google that invisibly tags content at the pixel level so that AI-generated or AI-edited media can be identified. According to Apple, this watermarking doesn’t visibly alter images and is designed to survive compression, cropping, filters, and other common modifications.
Which Edits Trigger Watermarks, and How to Detect Them
Within the Photos app, tools like Extend and Reframe now generate artworks or edits with SynthID watermarks by default. For Clean Up edits, watermark inclusion depends on the size of the changes made—larger or more substantial edits tend to generate the watermark, while minor tweaks may not. Standard image generation using Image Playground also always includes SynthID watermarks.
Users looking to check whether an image has a SynthID watermark can’t do it directly on their iPhone. Instead, detection is available through Google’s official SynthID Detector website.
What SynthID Is, and Why It Matters
SynthID is Google’s system for watermarking AI-generated and AI-edited media without affecting their visual quality. The watermark is embedded into the image pixels and designed to persist when content is altered, shared, compressed, or cropped. It’s already being used by multiple companies besides Google, including OpenAI, Nvidia, and others.
Apple first announced its plans to adopt SynthID during its rollout of Apple Intelligence earlier this year. The feature was promised in a June 2026 announcement: any image adjusted using Apple Intelligence would automatically carry the hidden SynthID watermark, and generative features in Image Playground would similarly tag outputs with SynthID.
That means photos taken with a regular camera capture, without edits or AI involvement, won’t be watermarked under the new policy. The watermarking specifically applies when AI tools are used to either generate new content or modify existing images with noticeable changes. Smaller, localized edits may not trigger the mark.
Apple’s adoption of SynthID arrives against a backdrop of global regulatory changes. Since August 2026, under the European AI Act and similar legislation elsewhere, providers of generative AI models are required to mark AI outputs in a detectable way. Leveraging SynthID helps Apple meet these transparency standards.
Visible metadata, like EXIF, can be removed easily and often falls off when images are shared; invisible watermarking like SynthID offers a more durable method for attributing content origin. While not foolproof—very aggressive edits, extreme cropping or adversarial attacks can potentially remove or degrade the watermark—SynthID is currently considered robust against typical transformations.
Apple users who upgrade to iOS 27.2 will start seeing SynthID applied in relevant tools immediately, but there’s no built-in feature in Photos for inspecting the watermark. To verify if an image has one, using the external SynthID detector will be necessary.
What this means:The move signals a deeper shift in how Apple handles AI-generated content—steering toward greater accountability and traceability. As digital media proliferates, distinguishing between genuine and AI-modified imagery is growing more vital for trust, copyright, and misinformation concerns. Users should be aware that what they edit may soon bear an invisible signature. Regulators will likely pay close attention to implementation details like which tools trigger the watermark and how well it survives edits. Expect follow-up scrutiny on edge cases—cleaner UI date stamps, detection thresholds, and support across third-party apps.
