New transparency rules for artificial intelligence systems come into force in the EU
Kyiv • UNN
The EU has introduced mandatory labeling of AI-generated content and notification of interaction with AI systems. Violations are subject to fines of up to 15 million euros.
New rules on the transparency of artificial intelligence systems have come into force in the European Union. They provide for mandatory labeling of certain content created using AI and informing users when they are interacting not with a human, but with an artificial intelligence system. This is reported by the official website of the European Commission, writes UNN.
Artificial intelligence is developing rapidly, making it increasingly difficult to distinguish content created or modified using AI from materials created by humans and those that are authentic. This creates new risks of large-scale spread of disinformation and manipulation, fraud, impersonation, and misleading consumers
It is noted that the new transparency requirements will help people recognize when they are interacting with artificial intelligence or seeing content created with its help. This will allow them to make more informed decisions and better protect themselves from disinformation and deception
Transparency requirements for developers and users of certain AI systems
Labeling and marking of AI-generated content: certain materials created or modified using artificial intelligence must be clearly and prominently labeled and contain machine-readable markers. The European Union has developed a set of special icons that can be used for this purpose
This applies to:
- transparency during interaction with an AI system: users must be clearly informed if they are communicating not with a real person, but with an artificial intelligence system, such as a chatbot, AI agent, or avatar.
- the European Commission has published recommendations to help developers and users of AI systems comply with these transparency requirements. They explain how compliance with the rules can be demonstrated, including through adherence to a special code of practice.
Enforcement of rules and sanctions
Compliance with transparency requirements will be monitored by national market surveillance authorities, the European Artificial Intelligence Office (for systems under its supervision), and the European Data Protection Supervisor (if the developers or users of AI systems are EU institutions).
Violations are subject to fines:
- up to 15 million euros or 3% of global annual turnover for companies;
- up to 750 thousand euros for EU institutions, bodies, and agencies;
- taking into account the principle of proportionality for small and medium-sized enterprises and small mid-cap companies.
The EU Artificial Intelligence Act (AI Act) came into force on August 1, 2024. Its provisions are being implemented in stages, with different requirements taking effect at different times. The Act creates a single market and harmonized rules for trustworthy artificial intelligence in the EU, promoting the development and deployment of AI.
At the same time, the document aims to reduce potential risks to health, safety, and fundamental rights of people, as well as to protect democracy and the rule of law. Ensuring the effective enforcement of the law is currently one of the key priorities of the European Commission.