Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI) sets out a framework for the development, commercialisation, deployment and use of artificial intelligence (AI) systems in the European Union. Its main objective is to promote the development of human-centred and reliable AI while ensuring a high level of protection of the health, safety and fundamental rights of citizens. The Regulation applies to all AI systems used in the EU, including those developed outside the Union, if their results are used in the EU.
 
The Regulation prohibits certain practices related to AI:
1. Use of AI to manipulate behaviour - systems that substantially alter human behaviour in a way that may cause significant harm.
2. Using AI to exploit vulnerable groups of people - especially children, the elderly or vulnerable people.
3. Biometric categorisation - systems that classify people based on biometrics with the potential to discriminate.
4. Using AI for real-time remote biometric identification in public spaces - except in certain situations such as law enforcement.
 
According to Regulation (EU) 2024/1689, AI systems are classified into high-risk and general-purpose systems. High-risk AI systems include systems for critical infrastructure such as transport and energy, which can endanger life and safety, systems for education, employment and human resources, for law enforcement 1for example used for risk assessment and identification, for social and financial services - such as determining creditworthiness or access to welfare benefits. General-purpose AI is an AI system that is not developed for a specific task but can be applied to a wide range of domains and tasks. Such systems can be used in a variety of contexts, such as natural language processing, pattern recognition, process automation, etc. They are versatile and can be adapted to different applications, making them more flexible, but also requiring more stringent security and ethical measures.
 
High-risk systems must meet strict requirements, including for:

1. risk management - risk identification and assessment
2. transparency - user awareness of the use of AI
3. monitoring and documentation - continuous monitoring and maintenance of documentation of system compliance
4. accuracy, reliability and security - systems must be designed for maximum accuracy and safety
5. existence of mechanisms for human intervention if necessary.

The Regulation requires suppliers of AI systems to:

• Conformity assessment: ensuring that systems meet regulatory requirements before being placed on the market
• documentation and registration: maintaining technical documentation and registering the system in relevant databases
• monitoring and maintenance: monitoring the system after it has been placed on the market and taking corrective action where necessary.

Deployers of AI systems are expected to:

• correct use: use of the systems in accordance with the instructions and limitations provided by the suppliers
• user awareness: ensuring that users are informed about the use of AI and its capabilities and limitations
• monitoring and response: monitoring system performance and responding to incidents or deviations.

Due to the high complexity of high-risk AI systems and the associated risks, the Regulation introduces a procedure for assessing their compliance with the requirements of the Regulation by national regulatory authorities (notified bodies) is introduced. They shall carry out:

• Assessment and certification of compliance of high-risk AI systems with the requirements of the Regulation before they are placed on the market.
• Surveillance and audit of certified systems to ensure that they continue to meet regulatory requirements.
• Awareness and transparency of their assessments and certifications through public registers and reports.

 
AI regulatory sandbox will be established at national level to facilitate the development and testing of AI systems under strict regulatory oversight before these systems are commercialised or otherwise made operational, further by regulatory sandbox laboratories that are controlled environments for testing the systems. Ongoing post-market monitoring, registration and traceability are in place, as well as penalties for non-compliance with the Regulation.
 
The Artificial Intelligence Office (AI Office) is established as a central body within the European Union to coordinate and manage AI activities. Its main functions include:

1. Oversight and regulation: ensuring compliance with AI regulations
2. Advice: support and guidance to Member States and industry
3. Innovation coordination: Supporting the development and implementation of innovative AI technologies
4. Risk assessment: identification and management of risks related to AI systems in the EU.

The AI Office will play a key role in implementing the Regulations and promoting ethical and safe AI practices in the EU.
 
Providers of non-high-risk AI systems are encouraged to develop codes of conduct designed to encourage the voluntary implementation of some or all of the mandatory requirements applicable to high-risk AI systems, appropriately adapted to the purpose of the systems and lower risk, considering available technical solutions and industry best practices. These will be developed with the involvement of different stakeholders and will be implemented voluntarily but encouraged through recognition by regulatory authorities.
 
Regulation (EU) 2024/1689 should apply from 2 August 2026. At the same time, given the unacceptable risk associated with the use of AI in certain ways, the prohibitions as well as the general provisions of this Regulation should apply from 2 February 2025. The AI Office should ensure that classification rules and procedures are up-to-date in the light of technological developments. In addition, Member States should lay down and communicate to the Commission the rules on penalties, including administrative fines and pecuniary sanctions, and ensure that they are correctly and effectively implemented by the date of application of this Regulation. The provisions on penalties should therefore apply from 2 August 2025.
 
Regulation (EU) 2024/1689 establishes a framework for the management, control and promotion of innovation in the field of artificial intelligence in the EU. Through its strict requirements and support measures, it aims to ensure the safety, transparency and ethics of AI systems, while stimulating innovation and competitiveness in the European market.
 
Regulation (EU) 2024/1689 - EN - EUR-Lex (europa.eu)