Artificial Intelligence regulatory initiatives of EU countries

AI regulatory initiatives of EU countries

On April 21, 2021, the EU Commission adopted a proposal for a regulation on artificial intelligence  (Artificial lntelligence Act). Regulation describes the first legal framework on AI in EU. It is expected that the AI Act will impose significant obligations impacting businesses across many, if not all, sectors of the economy.  Now the AI Act  goes through the legislative process and will enter into force in 2022 (expected). 

Meanwhile the European Commission and OECD published a report AI Watch - National strategies on Artificial Intelligence: A European perspective, 2021 edition. All EU Member States are developing and implementing policies and national strategies to seize the benefits of AI for the economy and society. Publication provides information on the latest policy developments and trends in members AI policies across the following policy areas:

The use of certain AI technologies raises ethical and legal issues, necessitating the development of a legal framework for AI.

Based on AI Watch, we would like to present the main internal regulatory initiatives of EU countries that are already in place showing what areas regarding AI are valuable to them. We see that EU countries identified and found important following areas which they decided to regulate (even partially).

1. AI Ethics

Many EU countries appointed AI ethics and knowledge centres to create places for experts to help resolve legal issues and matters of ethics relating to AI.

Additionally some countries have already implemented regulations directly or indirectly regarding AI, for example:

• Denmark proposes a specific focus on the responsible and sustainable use of data by the public and private sector. Danish government adopted a law on data ethics compliance for Denmark’s largest companies. The law requires companies with a data ethics policy to provide information on compliance, while companies without a data ethics policy must explain why they do not have a policy.
• Finland released the government report on Ethical information policy in an age of artificial intelligence outlining principles for fair data governance, including guidelines for the use of information and ethical values. 
• Malta has developed an AI certification framework, issued by the Malta Digital Innovation Authority (MDIA). The aim is to recognise that the AI systems have been developed in an ethical, transparent and socially responsible manner.
• Helsinki and Amsterdam have launched open AI registers that track how algorithms are being used in the municipalities.

2. Automated vehicles

Some countries have adopted regulations (including guidelines) to allow for the testing of automated vehicles and technologies on public roads.

• Belgian government adopted a regulation on tests with automated vehicles. It paves the way for the development of fully autonomous vehicles, as it allows the use of automated vehicles on the road for testing purposes and under restricted conditions.
• Finland adopted the Road Traffic Act to improve the smooth running and safety of transport and create preconditions for the digitalisation and safe automation of traffic.
• Lithuania adopted a law on autonomous driving that allows operation of self-driving cars without a driver being present.
• Netherlands adopted Experimental Law on self-driving vehicles for public road tests involving self-driving vehicles without drivers.
• Germany prepared ethical guidelines for self-driving cars.

3. AI and the healthcare

Other regulatory fields that receive particular attention are health care, data protection and automated decision making. Norway is working on proposals for amending its Health Register Act to delineate the use of data for patient treatments and rules of consent from individuals. They are working on amendments to the Health Personnel Act regarding the Duty of confidentiality and the right of disclosure in order to enable the use of algorithms for decision support on health data. According to the AI Watch Report this is still ongoing.

In Switzerland the Federal Office of Public Health (FOPH) monitors the impact of AI on medicine to include potential revisions to the existing legislation for data protection and privacy, and on the Federal Act on Medicinal Products and Medical Devices for the use of AI in the clinic process.

4. AI Sandboxes

Some governments are also considering the establishment of regulatory sandboxes. The objective is to facilitate experimentation in real-life conditions while temporarily reducing regulatory burdens to help testing AI. However development of sandboxes is still at an embryonic stage in most countries.

5. Specific regulatory initiatives adopted in Netherlands

• principles of good governance and consumer protection for the use and development of AI have been amended in the Freedom of Information Act, the General Administrative Law Act and the General Data Protection Regulation (GDPR); 
• legislation containing provisions regarding automated decision making for law enforcement;
• law preventing discrimination in recruitment when automated systems are used;
• guidelines for applying algorithmic data analysis, with the intention to translate them into legislation.

Summary

AI Watch Report shows that there is no one-size-fits all solution and there is a need for development of sector-specific regulatory frameworks, ensuring that the regulatory needs for AI development are adapted to the relevant industry areas. Additionally it is important to develop legislation that is technology-neutral where possible. 

Publication shows that scandinavian and Benelux countries are at the fore in regulations with relevant high degree of advancement, which also results from the high progress in implementing AI in various areas of life.

Most of countries commits to establish a regulatory basis for the development and use of trustworthy AI in line with international regulatory and ethical standards. Therofore EU wide Artificial Intelligence Act is needed and expected to build the grounds for further development of local and sector-specific regulatory frameworks.