The Policies of the EU AI Act in a Overview

In March 2024, the European Union adopted the AI Act, which regulates the use of AI in the EU. The act aims to protect fundamental rights, democracy and environmental sustainability from high-risk AI systems and to promote innovation.

The AI Act is the world’s first attempt at concrete regulation of AI. The objective is to promote the creation of a digital economic area between EU member states. Here is an overview of the most important points:

Act on the Regulation of Artificial Intelligence: A Guide

Part 1: Expectations and implications

The development of AI systems is progressing rapidly. The AI Act can be seen as a legal experiment because the way in which the law regulates something new is not already prescribed [1].

• Expectations: The AI systems used in the EU must be safe, transparent, comprehensible, non-discriminatory and environmentally friendly. These systems should be monitored by humans, not by automation, to prevent harmful events. [2]
• Impact on small businesses: The law is intended to give small businesses and start-ups the opportunity to develop and train AI models before they are made available to the public. [3]

Part 2: Requirements and sanctions

However, the regulation of artificial intelligence is a step in the right direction to give companies a direction of action and behaviour in relation to new AI systems. This will ensure transparency and strengthen trust in the use of technical innovations.

• Requirements: Member states must set up real-world laboratories and simulate realistic conditions to support the development of AI systems [3].
• Non-compliance: Companies that violate the provisions of the AI law can be fined, the amount of which depends on the extent of the violation and the size of the company [3].

AI categories and regulatory approaches

The AI Act bans AI systems with unacceptable risk in Europe. Targeted misrepresentation and misreporting designed by artificial intelligence can destabilise societies. Based on images and voice recordings, AI can generate fake material that could influence political elections, for example. The regulation of AI is intended to minimise this type of threat. [1]

Part 3: Risk-based approach

Prohibited AI systems with unacceptable risk

AI systems are banned if they are considered a threat to humans, including the following technologies [2]:

• Cognitive behavioural manipulation
• Social scoring
• Biometric identification of people
• Real-time biometric remote identification systems

An exception is made solely for law enforcement purposes in relation to real-time remote biometric identification systems. In principle, their use is also prohibited in law enforcement. However, there are detailed and narrowly defined exceptions in which real-time remote identification is permitted under strict security conditions. These systems may be used, for example, to search for a missing person or to prevent a terrorist attack. [2]

Monitored AI systems with high risk

AI systems that pose a high risk to the health and safety or fundamental rights of natural persons are considered high-risk (European Parliament, 2024). In order to increase its safety during use, the following regulations are issued, which must be complied with within the EU [4]

• Regular assessment of AI systems for accuracy, safety and fairness
• Risk reduction of the AI systems
• Automatic record oft all use cases of AI
• Be transparent
• Monitoring by humans

In addition, the public will in future have the right to lodge complaints about AI systems and demand an explanation if decisions are made on the basis of high-risk AI systems that affect their rights.

Regulated AI system with low risk

General purpose AI systems (e.g. generative foundation models such as ChatGPT) are categorised as low-risk, but must meet some transparency requirements [4]:

• Compliance with EU copyright law
• Prevent the creation of illegal content
• Publication of summaries of the content used for training
• AI-generated content must be labelled (e.g. deepfakes, audio and video content)

Why regulation of artificial intelligence is essential

‘Standardised regulation of the development, operation and use of AI is essential for the preservation of (competitive) fairness, human dignity and justice as pan-European principles.’
Oliver Bülchmann, Institute for Social Strategy

But what is the actual intention behind the legal regulation of AI?

The AI Act aims to minimise the risks of artificial intelligence, e.g. by protecting privacy, data protection and protection against discrimination, misrepresentation and false reports. At the same time, the opportunities of AI are to be promoted in order to create an innovation-friendly space within Europe. Productivity growth, efficiency gains and process optimisation in the areas of health, environmental protection, industry, mobility and care are just some of the opportunities made possible by AI. [1]

Timetable for implementation

Part 4: Timetable

The regulations will enter into force gradually, starting in September 2024 with the ban on AI systems with unacceptable risks and ending in March 2026 with the full application of all regulations [5].

The EU AI law is a significant step towards regulating and promoting the use of AI in the European Union. It aims to harness the potential of AI while minimising the risks in order to create an innovative and safe digital space.

---

Takeaway points:

• The EU AI Act regulates the use of AI and aims to protect fundamental rights, democracy and environmental sustainability.
• First global attempt at such regulation, promotes digital economic area.
• Key points of the act are the expectations of AI systems, effects on small companies, requirements and sanctions as well as various regulatory approaches.
• The regulation concerns AI systems with unacceptable, high and low risk, introducing strict rules and monitoring mechanisms.
• Aim of regulating AI: harnessing AI potential, minimising risk, creating an innovative and secure digital EU area

For more Information

EQUAL CARE certifies medical intervention with a balanced gender representation in data and evidence. Join EQUAL CARE today and lead the health market with our certification. Together, we can set a new standard for healthcare excellence and create a future where everyone receives the care they deserve.

Let’s meet: https://calendly.com/thao_equalcare/30min

If you would like to find out more about the topics of gender-specific medicine and the related work of EQUAL CARE,
visit us on Instagram, X, LinkedIn or on our website www.equal-care.org.

Sources:

[1] Bülchmann, O. (2023). Risiken begegnen, Potenziale entwickeln – KI-Technologie braucht Regeln und Freiheit gleichermaßen. Wirtschaftsinformatik & Management, 15(6), 420–430. https://doi.org/10.1365/s35764-023-00504-z

[2] KI-Gesetz: Erste Regulierung der künstlichen Intelligenz. (2023). Europäisches Parlament. https://www.europarl.europa.eu/topics/de/article/20230601STO93804/ki-gesetz-erste-regulierung-der-kunstlichen-intelligenz

[3] Artificial Intelligence Act: Deal on comprehensive rules for trustworthy AI. (2023). European Parliament. https://www.europarl.europa.eu/news/en/press-room/20231206IPR15699/artificial-intelligence-act-deal-on-comprehensive-rules-for-trustworthy-ai

[4] Gesetz über künstliche Intelligenz: Parlament verabschiedet wegweisende Regeln. (2024). Europäisches Parlament. https://www.europarl.europa.eu/news/de/press-room/20240308IPR19015/gesetz-uber-kunstliche-intelligenz-parlament-verabschiedet-wegweisende-regeln

[5] Artificial Intelligence Act: MEPs adopt landmark law. (2024, März 13). European Parliament. https://www.europarl.europa.eu/news/en/press-room/20240308IPR19015/artificial-intelligence-act-meps-adopt-landmark-law