On August 2, 2026, the European Union flips a switch that will change how AI-generated content circulates online. Under the EU AI Act, any synthetic image, video, audio clip, or piece of text designed to look authentic must carry a visible label and an embedded digital watermark. Companies that fail to comply face fines of up to 7 percent of their global annual turnover.
The rule sounds straightforward. In practice, the gap between what the law demands and what the technology can deliver is wide enough to drive a truck through. And for creators, developers, and anyone publishing AI-generated work, understanding that gap is the difference between compliance and a regulatory headache.
What the EU Rules Actually Require
The AI Act’s transparency obligations, which take effect for new AI systems on August 2 and for existing systems four months later, target content that is designed to look real. That includes photorealistic AI-generated images, synthetic audio that mimics a real person’s voice, AI-written text presented as factual reporting, and deepfake videos.
The requirements are two-pronged. Content must carry a visible label — something a human can see or hear that indicates the material is AI-generated. It must also contain a machine-readable watermark, invisible to the user but detectable by automated systems. The idea is that even if someone strips the visible label, the watermark survives, and platforms can flag the content downstream.
Sergey Lagodinsky, the Green MEP who helped negotiate the AI Act, framed it as a democracy issue. “Making transparent this information is something which we need to preserve our democracy and the authenticity of facts online,” he told the Guardian. The law is not just about consumer protection. It is about making it harder to flood the information ecosystem with synthetic content that passes for real.
Enter SynthID: The Watermark That Survives
Google’s SynthID is currently the most prominent technology trying to make watermarks stick. Unlike metadata-based approaches that can be stripped by simply screenshotting an image or re-encoding a video, SynthID embeds its signal directly into the pixels of an image or the waveform of audio. The watermark is imperceptible to humans but detectable by a specialized decoder — and it survives cropping, compression, resizing, and even screenshots.
Here is how it works under the hood. SynthID modifies the pixel values of an image in a pattern that is statistically indistinguishable from random noise to the human eye but forms a recognizable signature to the decoder. For images, the watermark is embedded during the generation process itself — the AI model learns to produce images that contain the signal natively, rather than having a watermark applied as a post-processing step. For audio, it works similarly by shaping the waveform during synthesis. This baked-in approach is what makes SynthID harder to remove than a traditional overlaid watermark: you cannot strip it without fundamentally altering the image in ways that degrade quality.
The scale of the problem SynthID is trying to solve is hard to overstate. Starling Lab, a research collaboration from Stanford and USC, estimates that it took 149 years after the invention of the camera — from 1826 to 1975 — for humanity to create 1.5 billion images. Generative AI matched that number in 18 months. Google alone reports that more than 100 billion AI images and videos have been created using its tools in just a couple of years.
At its spring I/O conference, Google announced that OpenAI, Runway, Nvidia, and other major AI developers are adopting SynthID. The coalition matters because watermarking only works at scale if the major content-generation platforms all participate. If Stable Diffusion images carry a SynthID watermark but Midjourney images do not, the system fails at its most basic job: universal detection. The consortium approach — getting competitors to agree on a shared standard — is arguably more important than the technical sophistication of the watermark itself.
Ars Technica’s Ryan Whitwam tested SynthID extensively and found that it holds up under real-world abuse. Screenshots, heavy JPEG compression, minor edits — the watermark survives all of it. Whitwam took watermarked images, ran them through aggressive compression pipelines, resized them, cropped them, and re-encoded them in different formats. The decoder still picked up the signal in most cases. It is not unbreakable. Determined adversaries with enough technical skill can likely strip or degrade the signal through adversarial perturbations or by regenerating the image through another AI model. But for casual attempts to circumvent labeling — the kind that account for most real-world misuse — SynthID is genuinely robust.
The Two Standards Problem
There is a second approach to AI content labeling, and it works completely differently. C2PA, or the Coalition for Content Provenance and Authenticity, is an Adobe-led standard that attaches cryptographically signed metadata to media files. When you generate an image in Adobe Firefly or edit a photo in Photoshop with Content Credentials enabled, the file carries a tamper-evident record of its origin: where it came from, what AI tools were used, and what edits were made.
C2PA and SynthID are not competitors so much as answers to different questions. C2PA answers “where did this file come from?” It works great when the metadata chain remains intact — when a file is shared on platforms that support C2PA and viewed in software that can read the credentials. It breaks when someone takes a screenshot, which discards all metadata.
SynthID answers “is this content AI-generated?” It survives screenshots but provides no provenance information — no record of which tool created the image or when. A combined approach, where SynthID handles detection and C2PA handles attribution, is the direction the industry is heading.
What the Rules Actually Miss
The EU AI Act’s labeling requirements face two problems that no watermarking technology can solve.
First, the rules only apply to content generated by compliant systems. A bad actor who wants to create a deepfake of a politician does not use Google’s Imagen with SynthID enabled. They use an open-source model running on a local GPU, or a tool hosted in a jurisdiction that does not enforce the EU rules. The labeling requirement creates a compliance burden for legitimate creators while doing nothing to stop intentional deception. The people who are most likely to misuse AI-generated content for disinformation are also the least likely to use tools that voluntarily embed watermarks.
Second, detection is only as good as the platforms that enforce it. SynthID requires a decoder to check for the watermark. If social media platforms do not run that decoder on every uploaded image — and they currently do not — a watermarked image and an unwatermarked one are indistinguishable to the average user scrolling through a feed. The EU rules mandate that content be labeled at the point of creation, but they do not mandate that platforms verify those labels or check for watermarks at the point of distribution. That gap between creation and distribution is where most disinformation does its damage.
There is a third issue that gets less attention: watermarking works against the grain of how creative professionals actually use AI tools. An artist who uses AI to generate a base image, then spends hours painting over it in Photoshop, has created something that is neither purely AI-generated nor purely human-made. SynthID can detect the AI component even after heavy editing. But should that hybrid work be labeled as AI-generated? The law does not have a clear answer, and the ambiguity puts creators in a difficult position.
What Creators and Developers Should Do Now
If you publish AI-generated images, video, or audio — especially if you have users or audiences in the EU — here is the practical starting point.
Use tools that support SynthID or C2PA. Google’s Imagen, Adobe Firefly, and tools built on top of OpenAI’s and Runway’s APIs are increasingly adopting these standards. If your workflow involves open-source models like Stable Diffusion or Flux, you will need to add labeling yourself, either through post-processing tools or by selecting a distribution platform that applies watermarks automatically. Some hosting services and CDNs are beginning to offer automatic C2PA signing as a feature — worth checking if your platform supports it.
Label visibly even when the law does not require it. The EU rules are the floor, not the ceiling. Visible labels — a “Generated with AI” badge or a note in the description — build trust with audiences and reduce the risk of your work being misused or misinterpreted. They also future-proof you against regulations that are likely to tighten over time. Several major stock photo platforms, including Shutterstock and Adobe Stock, already require visible AI labels alongside whatever invisible watermarks they embed. That pattern is likely to spread.
Audit your pipeline for gaps. If you use multiple AI tools in a workflow — say, Midjourney for initial generation, Photoshop for editing, and Topaz for upscaling — check whether each tool preserves or strips watermarks from the previous step. A watermark embedded by Midjourney is useless if your upscaler obliterates it. This kind of pipeline audit is tedious but it is the only way to know whether your compliance measures actually survive through to the final output.
Assume the technical solutions are incomplete. Watermarks can be stripped. Metadata can be discarded. Even SynthID, as robust as it is, is not a guarantee. Treat labeling as one layer of a broader approach to transparency — alongside clear disclosures, audience education, and responsible publishing practices. The best defense against AI disinformation is not any single technology. It is a combination of tools, norms, and a public that knows to ask where content came from.
The Bottom Line
The EU’s labeling rules are the most significant AI transparency regulation to date, and they will force real changes in how AI-generated content is created and distributed. Google’s SynthID and the C2PA standard provide workable technical tools for compliance. But the gap between regulatory aspiration and technical reality remains wide, and closing it will require platforms, tool developers, and creators to do more than the minimum the law requires.
For now, the practical advice is boring but correct: label your AI-generated content visibly, use tools that embed watermarks when you can, and do not assume that a watermark makes a deepfake problem go away. The rules are a start. They are not a solution.