EU AI Act: Mandatory Labeling of AI Content and Chatbots Takes Effect

On 2 August 2026, the European Union brought the transparency obligations of the AI Act into force: companies must now disclose when users are interacting with AI systems and label content generated or altered by neural networks.

What has changed for companies?

The new requirements split responsibility between two types of market participants. Providers — companies that develop AI systems and place them on the market — must build disclosure mechanisms directly into their products' architecture. Deployers — platforms and services that use ready-made AI systems — must explicitly notify end users that they are interacting with AI.

The line between the two roles is not always clear. A number of large players, Meta and SpaceXAI among them, both develop their own AI systems and embed them in their platforms, so they bear responsibility as both providers and deployers.

The distinction matters: the provider is responsible for making sure the system can disclose its nature; the deployer is responsible for making sure that disclosure actually happens at the point of contact with the user.

What exactly must be labeled?

The new requirements cover chatbots, deepfakes and any synthetic content. According to the European Commission, the key parameters are as follows:

• Effective date — 2 August 2026
• Chatbots must clearly tell the user that they are talking to AI, not a human
• Synthetic content — images, video and audio generated or altered by AI — must be labeled accordingly
• Deepfakes fall under the same disclosure requirements
• Companies that combine both roles (such as Meta and SpaceXAI) carry cumulative obligations

The European Commission has developed a standard set of AI icons — a ready-made visual language that platforms can use instead of creating their own markers. The logic is simple: if enough services adopt the same icons, users will learn to recognize them faster.

"Safer and more transparent AI" — this is how the

European Commission described the goal of the new package of requirements in its official press release of 2 August 2026.

Why does this matter outside the EU?

De jure, the AI Act applies only to Europe, but de facto it affects any platform that serves European users. For international services, this means concrete compliance tasks: revising UX notifications, auditing scenarios in which users interact with AI features, and integrating the standard labeling. According to European Commission materials, the AI Act is the world's first comprehensive legal instrument of this scale in the field of AI.

Failure to comply with the new requirements carries penalties under the enforcement mechanisms that EU regulators are building in parallel with the Act itself. Similar initiatives for labeling AI content are under discussion in the US, the UK and several countries in Asia, so the practice introduced in Europe risks becoming the de facto global standard.

What this means

The EU AI Act introduces the first mandatory, large-scale regime for labeling AI content in a major region. For users, it offers a minimal safeguard against hidden chatbots and unlabeled deepfakes. For the industry, it is a signal that regulatory pressure for AI transparency will only grow.

Frequently asked questions

When did the EU AI

Act's transparency requirements take effect?

The transparency requirements took effect on 2 August 2026 — a date set out in the European Commission's official documents.

How do the obligations of providers and deployers differ?

The provider — the company that built the AI system — must design it with mechanisms that clearly disclose its nature. The deployer — the platform using the system — is responsible for making sure the user actually receives that disclosure at the moment of interaction. Companies like Meta and SpaceXAI are both providers and deployers at once and carry both types of obligations.

*Meta is recognized as an extremist organization and banned in Russia.