WhatsApp is rolling out a new feature that permanently labels AI-generated media. With this move, the messenger implements a core requirement of the EU AI Act—and makes transparency a mass-market reality. The function is available to channel administrators via context menu and applies to images, videos, audio clips, and GIFs. Crucially: once set, the label cannot be removed.
The essentials
- Channel administrators can permanently mark AI content with a label; followers see the origin permanently
- Covers images, videos, audio, and GIFs from all major AI services – Meta AI, ChatGPT, Gemini, Midjourney, DALL-E
- Text messages remain exempt for now
- Rollout: Android first, iOS version planned
- EU AI Act: As of August 2, 2026, Regulation EU 2024/1689 requires clear labeling of AI-generated content
Compliance meets the mainstream
The timing is no coincidence. The EU AI Act makes labeling AI-generated content a legal requirement—a tool against disinformation. WhatsApp is anticipating this requirement and building it directly into the platform. That's remarkable: not just enterprises must adapt to new rules, but millions of users and administrators will experience compliance requirements firsthand.
The labeling applies across platforms—whether the content comes from Meta's own tools or competitors like OpenAI or Google. The message is clear: this is about transparency, not market protection.
Beyond labels: New admin tools
Meanwhile, WhatsApp is expanding management capabilities. Groups with more than 32 members get an @all mention that notifies all participants simultaneously. Administrators can set deadlines for polls and enable anonymous voting. New is also the ability to create a similar group with an existing member list in one click.
For Apple users, WhatsApp simplifies child account setup: QR codes and invite links connect child accounts directly to parental supervision—a step toward family compliance.
External AI integration accelerates
The news also shows: WhatsApp is becoming a platform for external AI providers. Developers found code in the ChatGPT app for Android for a WhatsApp sticker export—users could turn AI images directly into sticker packs with three to 30 designs. Perplexity AI is integrating the Gemini-2.5-Flash model into its WhatsApp bot, enabling image editing via text command directly in chat.
This means: AI-generated content won't just be labeled in the messenger—it will also be created there. The labeling function becomes an immediate necessity.
What this means for German enterprises
For compliance officers in Germany, this is a signal: labeling AI-generated content isn't just a concern from August 2026 onward—it's already here. Anyone running or planning channels on WhatsApp should familiarize themselves with the new tools now. At the same time, it's clear: the EU AI Act won't only be enforced by authorities but embedded by platforms themselves at scale. This reduces uncertainty but raises questions: How will labels be used in practice? Will they become routine or exception? And is labeling enough to combat disinformation—or do additional measures matter?
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
