• by Laurie Sullivan , Staff Writer, Yesterday

The European Commission's new transparency rules for AI systems took effect on August 2, 2026, mandating clear disclosure when users interact with artificial intelligence (AI) or view artificially generated synthetic content for images, audio and video.

As of Sunday, the European Commission can begin implementing fines on companies that fail to comply with the act.

"We are taking an important step towards AI that people and businesses can understand and trust, and whose benefits are widely shared across our society," Henna Virkkunen, executive vice president for tech sovereignty, security and democracy of the European Commission, wrote in a X post on Friday. 

The regulations under Article 50 of the AI Act require providers to label synthetic media and notify individuals when "emotion recognition" or "biometric categorization systems" are being used.

The European Commission published guidelines to help "providers" and "deployers" of AI systems to meet transparency obligations.

The guidelines explain how compliance can be demonstrated, including through adherence to a code of practice.

Providers are defined as any person or company that develops an AI system, whereas deployers of AI systems are defined as any person or organization using an AI system for professional or commercial use.

Virkkunen explained that under the new rules, chatbots and other interactive AI systems will have to tell users they are using AI.

Deepfakes will require labels to identify the synthetic content, and AI-generated or altered content will need to have a machine-readable marks so it can be detected.

In June, retail lobby group Eurocommerce wrote a letter to Virkkunen sharing concerns about the guidelines.

"We strongly support the objective of enhancing transparency where AI-generated or AI-manipulated content is deceptive by creating, risks of harmful impersonation, disinformation, manipulation or fraud, which is essential to maintaining consumer trust and supporting the responsible uptake of AI in Europe," per the letter. "However, we believe, amongst others, that the current wide interpretation of what constitutes a deep fake risk undermining the objective of meaningful transparency by capturing non-deceptive commercial content."

Online travel company Booking.com is supportive of the AI Act, but told the Financial Times that being transparent about when consumers talk with or interact with AI leads to "better customer service."

Supporting the Act is one thing, but Booking.com wants more "clarity from the Commission on how to translate the legislation into its products," FT wrote, adding that an executive for the travel agency said the company needs to "understand in plain and simple terms what the regulation allows" so its engineers can better understand the nuances of the wording.

But the transparency rules are only one part of the AI act’s rollout. FT pointed out that the Commission also gained new legal powers to inspect general-purpose AI models that pose “systemic risks," something similar to what Trump proposed for the United States in May. 

If in breach of the AI act, the EU Commission can fine technology companies up to 3% their total worldwide annual turnover in the preceding financial year, or €15 million, whichever is higher, according to the policy.