EU AI Act Article 50: New AI transparency rules

The EU has also provided some helpful guidance on the implementation of the new obligations here.

Under Article 50, providers of AI systems must now:

1. ensure that AI systems intended to interact directly with natural persons are designed and developed in such a way that the natural persons concerned are informed that they are interacting with an AI system, unless this is obvious from the point of view of a natural person who is reasonably well-informed, observant and circumspect, taking into account the circumstances and the context of use; and
2. where such systems (including general-purpose AI systems) generate synthetic audio, image, video or text content, ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated. Providers shall ensure their technical solutions are effective, interoperable, robust and reliable as far as this is technically feasible, taking into account the specificities and limitations of various types of content, the costs of implementation and the generally acknowledged state of the art, as may be reflected in relevant technical standards. This obligation does not apply to the extent the AI systems perform an assistive function for standard editing or do not substantially alter the input data provided by the deployer or the semantics thereof.

There is a grace period (until 2 December 2026) for the marking and detection of AI-generated content of AI systems placed on the market before 2 August 2026.

Deployers (natural or legal persons using an AI system) must now:

1. in relation to an emotion recognition system or a biometric categorisation system, inform the natural persons exposed thereto of the operation of the system, and shall process the personal data in accordance with the GDPR and related EU legislation;
2. disclose that the following has been artificially generated or manipulated by an AI system, in particular:
   
   1. in relation to any image, audio or video content constituting a deep fake. However, where the content forms part of an evidently artistic, creative, satirical, fictional or analogous work or programme, the transparency obligations are limited to disclosure of the existence of such generated or manipulated content in an appropriate manner that does not hamper the display or enjoyment of the work; and
   2. regarding text which is published with the purpose of informing the public on matters of public interest. This obligation does not apply where the AI-generated content has undergone a process of human review or editorial control and where a natural or legal person holds editorial responsibility for the publication of the content.

The obligations above, in relation to both providers and deployers, do not apply where the use of such AI systems is authorised by law to detect, prevent, investigate or prosecute criminal offences. In the case of emotion recognition or biometric categorisation systems, such use shall be subject to appropriate third-party rights and freedoms safeguards.

Additionally, these obligations do not replace, limit or affect the requirements and obligations set out in Chapter III of the Act, which provides requirements, obligations and other rules relating to high-risk AI systems. This means that an AI system could simultaneously fall within the scope and obligations of this Article 50, and Articles 6-49 of the Act (pertaining to Chapter III), if such system is classified as high-risk.

Comments

The practical reach of Article 50 is both technologically and commercially broad. It applies to a wide range of popular AI tools, including chatbots, generative AI systems and deepfake technologies. It also captures both providers of AI systems and “deployers” (i.e. users) of such systems.

From a broader commercial perspective, the disclosure obligation at point 4 is particularly noteworthy. Use of AI to generate marketing collateral, including audiovisual assets and copy, has exploded. AI is also used more and more frequently in the generation of synthetic “influencers”. Anyone using AI for such purposes will now need to be transparent as to such use, or risk a significant fine.

These transparency obligations may also have significant implications from an IP perspective:

• Article 50 will require AI-generated or manipulated content to be marked and for the involvement of AI in its creation to be disclosed. In a world in which AI-generated inventions struggle for patent protection (see DABUS case), and AI-generated works are refused copyright, businesses using AI for value generation will need to consider how to protect that value.
• Litigants will be aware that their counterparties may have troves of information about the generation of the work being litigated, likely including metadata, watermarking records, prompt histories, generation logs and other provenance information. Even where works are not AI-manipulated, their owners may maintain records to evidence that fact. That information will be ripe for disclosure or discovery during litigation.
• Protection of personal image has long been a concern for individuals, which has been enhanced with the advent of AI image, video and audio generation. While Article 50 does not prevent deepfakes, it assists in guarding against them; deepfakes under the scope of Article 50 will have to be labelled as such, and if they are not, then their creators risk regulatory consequences.

The obligation upon AI providers to create an effective, interoperable, robust and reliable means of tracking AI creation or manipulation will create significant industry discussion. The Content Authenticity Initiative (Content Authenticity Initiative), led by Adobe, and similar projects are likely to become more influential in the face of Article 50.

While Article 50 does not create new rights, its transparency obligations are far-reaching and will add to the growing EU toolkit of rights protection in the face of AI.

All businesses, regardless of whether they are currently involved in litigation, should ensure that they are aware of, and in compliance with, the obligations under Article 50, given their geographically and commercially far-reaching nature.