Brussels,

THE NEW FRONTIERS

EU Artificial intelligence Act
​

Introduction

​
On 21 April 2021, the European Commission released a Proposal for a Regulation Laying down harmonized rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union legislative Acts.

The proposed regulation aims to establish a comprehensive framework for the development and deployment of artificial intelligence (AI) within EU. It addresses various aspects of AI, including ethical considerations, transparency, accountability, and legal implications. The regulation recognizes the potential benefits of AI while acknowledging the need to ensure its safe and responsible use.

The document outlines the key provisions of the proposed regulation.

It introduces requirements for high-risk AI systems, defined as systems that pose significant risks to health, safety, or fundamental rights. These requirements include mandatory conformity assessments, data and record-keeping obligations, and the appointment of a responsible person within the organization. The proposed regulation also emphasizes the importance of human oversight and the need for clear and understandable information for users.

Additionally, the document discusses the establishment of a European Artificial Intelligence Board and a European Artificial Intelligence Committee to facilitate cooperation and coordination among member states. It highlights the role of national competent authorities in enforcing the regulation and imposing sanctions for non-compliance.

The proposed regulation aims to strike a balance between promoting innovation and protecting individuals' rights and safety. It acknowledges the importance of fostering a trustworthy and human-centric approach to AI. The document also emphasizes the need for international cooperation and harmonization of AI regulations to ensure a global ethical and legal framework for AI development.

The proposed EU Regulation seeks to address the challenges and risks associated with AI while harnessing its potential for societal and economic benefits. It sets out a comprehensive framework to govern the development, deployment, and use of AI systems within the EU, with a particular focus on high-risk applications.
​

The Objectives of the EU Artificial Intelligence Act

  • ensure that AI systems placed on the Union market and used are safe and respect existing law on fundamental rights and Union values
  • ensure legal certainty to facilitate investment and innovation in AI
  • enhance governance and effective enforcement of existing law on fundamental rights and safety requirements applicable to AI systems
  • facilitate the development of a single market for lawful, safe and trustworthy AI applications and prevent market fragmentation.

Classification of A.I. level of risk 

 The drat of the Commission proposes to adopt different set of rules tailored on a risk-based approach with four levels of risks:
​
  • Unacceptable risk AI.  Harmful uses of AI that contravene EU values (such as social scoring by governments) will be banned because of the unacceptable risk they create; 
  • High-risk AI. A number of AI systems (listed in an Annex) that are creating adverse impact on people's safety or their fundamental rights are considered to be high-risk. In order to ensure trust and consistent high level of protection of safety and fundamental rights, a range of mandatory requirements (including a conformity assessment) would apply to all high-risks systems; 
  • Limited risk AI. Some AI systems will be subject to a limited set of obligations (e.g. transparency); 
  • Minimal risk AI. All other AI systems can be developed and used in the EU without additional legal obligations than existing legislation.

Commission's proposal: Critical points

1. Definition and Scope

The document provides a broad definition of AI systems and categorizes certain systems as "high-risk". However, the scope and criteria for determining which AI systems fall under this category may be subject to interpretation and require further clarification. Clearer guidelines and thresholds could help in ensuring consistent application and effective regulation.
  • Article 3: Definitions
  • Article 5: Scope


2. Proportionality and Impact on Innovation
  
While the proposed regulation aims to strike a balance between regulation and innovation, there is a concern about the potential burden it may impose on businesses, particularly small and medium-sized enterprises. The requirements and compliance obligations for high-risk AI systems might disproportionately impact smaller organizations, hindering their ability to compete and innovate in the AI market.
  • Article 44: Obligations for high-risk AI systems
  • Article 48: Conformity assessment procedures
  • Article 51: Conformity assessment bodies


3. Enforcement and Implementation Challenges

The proposed regulation places significant responsibility on national competent authorities for enforcement and supervision. Ensuring consistent implementation and enforcement across EU member states may pose challenges, particularly considering the varying levels of resources and expertise available to different authorities.
Harmonization and cooperation mechanisms need to be well-established to address these potential issues.
  • Article 61: National competent authorities
  • Article 62: Cooperation mechanism
  • Article 66: Powers of national competent authorities



4. Impact on International Competitiveness
​
The proposed regulation may impact the EU's competitiveness in the global AI market. Stricter regulations within the EU, if not matched by other regions, could potentially result in the relocation of AI development and deployment activities outside the EU. Balancing regulation with strategies to promote innovation and maintain competitiveness on a global scale will be essential.
  • Article 1: Objectives and general principles
  • Article 2: Definitions
  • Article 54: Cooperation with third countries
​

5. Technological Neutrality
 
While the proposed regulation acknowledges the need for a risk-based approach, there is a possibility that it could inadvertently favor or hinder specific AI technologies or approaches. Striking the right balance and ensuring technological neutrality will be crucial to avoid stifling certain AI developments or favoring specific solutions over others.
  •  Article 4: Risk-based approach
  •  Article 44: Obligations for high-risk AI systems
  •  Article 52: Transparency obligations


Picture
Click to read
Picture
Click to read

EU Council Position

The Council has adopted its common position (‘general approach’) on the AI Act on 6 December 2022. The Council’s text, inter alia:
​
  • narrows down the definition to systems developed through machine learning approaches and logic- and knowledge-based approaches;
  • extends to private actors the prohibition on using AI for social scoring;
  • adds a horizontal layer on top of the high-risk classification to ensure that AI systems that are not likely to cause serious fundamental rights violations or other significant risks are not captured;
  • clarifies the requirements for high-risk AI systems;
  • adds new provisions to account of situations where AI systems can be used for many different purposes (general purpose AI);
  • clarifies the scope of the AI act (e.g. explicit exclusion of national security, defence, and military purposes from the scope of the AI Act) and provisions relating to law enforcement authorities;
  • simplifies the compliance framework for the AI Act;
  • adds new provisions to increase transparency and allow users' complaints;
  • substantially modifies the provisions concerning measures in support of innovation (e.g. AI regulatory sandboxes).

European Parliament: its negotiation position

In Parliament, the discussions are led by the Committee on Internal Market and Consumer Protection (IMCO; rapporteur: Brando Benifei, S&D, Italy) and the Committee on Civil Liberties, Justice and Home Affairs (LIBE; rapporteur:  Dragos Tudorache, Renew, Romania) under a joint committee procedure. Parliament adopted its negotiating position (499 votes in favour, 28 against and 93 abstentions) on June 2023 with substantial amendments to the Commission's text, including inter alia,

1. Definition of AI Systems: Members of the European Parliament (MEPs) amended the definition of AI systems to align it with the definition agreed by the OECD.

2. Prohibited AI Systems: MEPs made substantial amendments to the list of AI systems prohibited in the EU. They proposed banning the use of biometric identification systems for both real-time and ex-post use, with exceptions for severe crimes and pre-judicial authorization. They also called for a ban on biometric categorization systems using sensitive characteristics, predictive policing systems, emotion recognition systems, and AI systems that scrape biometric data indiscriminately to create facial recognition databases.

3. High-Risk Systems: While the Commission proposed automatically categorizing certain systems as high-risk, Parliament added the requirement that systems must pose a "significant risk" to qualify. Deployers of high-risk systems in the EU would be required to conduct a fundamental rights impact assessment, including consultation with relevant authorities and stakeholders.

4. Regulation of General-Purpose AI Systems: Parliament aims to establish a layered approach to regulate general-purpose AI systems. Providers of foundation models would be obligated to protect fundamental rights, health, safety, the environment, democracy, and the rule of law. Generative foundation AI models, including those used for generating art, music, and other content, would be subject to stringent transparency obligations. Foundation models should provide necessary information to downstream providers for compliance with the AI Act.

5. Strengthened National Authorities and AI Office: National authorities' competences would be strengthened, and Parliament proposes the establishment of an AI Office, a new EU body to support the harmonized application of the AI Act, provide guidance, and coordinate cross-border investigations.

6. Exemptions for Research and Open-Source AI: To support innovation, Parliament agrees to largely exempt research activities and the development of free and open-source AI components from compliance with the AI Act rules.
Learn more about Parliament’s work on AI and its vision for AI’s future

Negotiations between legislators

EU lawmakers have started  the negotiations to finalise the new legislation. A first trilogue took place in June 2023. A second trilogue is scheduled to take place in July 2023. 

The EU investments on AI

  • HORIZON EUROPE
  • ​DIGITAL EUROPE PROGRAMME
<
>

Cluster 4: Digital, Industry and Space
​

The financing on Artificial Intelligence is inscribed in Cluster 4 of Horizon Europe Framework programme 2021-2027.
​​The overarching vision behind the investments under Cluster 4 is that of Europe shaping competitive and trusted technologies for a European industry with global leadership in key areas, enabling production and consumption to respect the boundaries of our planet, and maximising the benefits for all parts of society in the variety of social, economic and territorial contexts in Europe.
This contributes to a competitive, digital, low-carbon and circular industry, ensures sustainable supply of raw materials, develops advanced materials and provides the basis for advances and innovation in  global challenges to society.

Digital Europe programme
​

The Digital Europe Programme (DIGITAL) is a EU funding programme focused on bringing digital technology to businesses, citizens and public administrations.

The Digital Europe Programme, with a budget of €7.5 billion (2021-2027), and €1.3 billion for 2023-2024, aims to address challenges in supercomputing, AI, cybersecurity, digital skills, and the adoption of digital technologies. The program aims to accelerate economic recovery and drive digital transformation across Europe, with a particular focus on supporting small and medium-sized enterprises. The funding will complement other EU programs, including the Horizon Europe program for research and innovation, the Connecting Europe Facility for digital infrastructure, and the Recovery and Resilience Facility. The Digital Europe Programme is part of the Multiannual Financial Framework 2021-2027 and will work in conjunction with other initiatives to support the development of a digital Europe.