1. The impact of the AI Act on the trans-national governance of AI systems
Artificial intelligence (AI) is a transformative technology which brings both risks and opportunities, necessitating careful consideration. Striking a balance requires proceeding with caution, therefore taking a calculated approach at present can establish a basis for future regulations. However, it is crucial to start laying those foundations without delay.
On the 14th of June the European Parliament voted to approve the EU Regulation on Artificial Intelligence, putting Brussels a step closer to shaping global standards for AI. Through this act, the EU seeks to regulate AI systems within the Union and address various challenges associated with AI, such as bias and discrimination in decision-making algorithms, privacy and security concerns with the potential benefits of AI and the traditional notions of accountability and responsibility. Clearly, its impact on trans-national governance of AI systems will depend on the implementation by EU member states and on its reception by third countries.[1]
As it was with the GDPR and its success in influencing international standards, the new Act may serve as a model for other countries developing their own AI regulations, potentially leading to a harmonized global framework for AI governance. Different countries and regions might consider the EU’s approach as a blueprint for their own AI regulations or strive to align their regulations with the EU’s standards. This alignment would aim to promote seamless cross-border trade and collaboration in the field of AI.
However, the regulation is still awaiting to be negotiated within the European Council. Moreover, the actual impact of the AI Act on trans-national governance will depend on its acceptance and adoption outside the EU, where the need of a global debate is necessary.
2. The reaction of laws after the spread of AI technologies in Europe and beyond
The emergence of new digital technologies led, both in Europe and outside the European territory, to the pursuit of new laws, answers, and regulations for those systems. Particularly, at European level, the technological process has responded through the drafting of the Act, which has a huge potential to boost capital and labour productivity, innovation, growth, and job creation. At the same time, several countries are taking steps to regulate AI systems, each with its own approach. Looking at the global big players, China, after having established the “Next Generation Artificial Intelligence Development Plan” in 2017, subsequently established two laws regulating specific AI applications in 2022 and while some other provisions are still at the draft stage, China has regional legislation on AI. In fact, on the 6th of September 2022, the Shenzhen government published China’s first city-level AI regulation. Behind all these acts and drafts, the intention to strengthen the government control over China’s technological progress clearly emerges. On the other hand, the United States has published non-binding blueprints and frameworks to address potential harm and manage AI risks in 2022, known as Blueprint for an AI Bill of Rights, which, together with the AI Risk Management Framework, is the core of the US strategy. In this regard, unlike the EU Act, this is a non-binding act which is aimed to minimize potential harm from AI systems. The US legislation seems to be, and likely to remain, a mosaic of federal approaches rather than a centralized strategy, which will be a step behind, for now, compared to the EU approach.[2]
Currently, there is no global consensus on AI regulation, as most countries are at different stages of development and are adopting varying approaches to AI. However, the EU strives to be the world’s first super-regulator in AI. Achieving a balance between innovation and curbing potential harms remains a challenge which poses the main question of fostering innovation while aiming to restrict potentially detrimental applications of AI.
3. The EU current and future legal framework in tackling trans-national challenges of AI
The initiatives proposed by the European Union regarding AI face a more challenging path compared to other countries. In fact, the AI Act requires stringent monitoring and disclosure, which intensifies as the level of risk associated with AI systems increases. Therefore, it could be argued that the European Union is currently leading the way in emerging AI legislation. However, before we can truly determine the effectiveness of this system, we must await the implementation of the regulation by the Member States. Moreover, it may be necessary to establish a dedicated regulatory authority and potentially intergovernmental treaties. Additionally, looking towards the future, an international body might be required not only to standardize rules that favor the common good while preserving innovation but also to ensure compliance and develop tools and expertise to study and manage the impacts of AI ex-ante. This could be similar to the European Union Agency for Cybersecurity (ENISA), which aims to achieve a high common level of cybersecurity across the Union in cooperation with the wider community.
4. Fostering the protection of fundamental human rights and constitutional principles vis-à-vis the rise of AI
The AI Act will play a crucial role in fostering the protection of fundamental human rights and constitutional principles in the context of the global rise of AI without stifling innovation.
Given that it applies to AI systems providers, importers, distributors and users, as well as product manufacturers, the Act, through its wide scope, prioritizes ethical and human-centric aspects of AI development and use, placing a key goal on securing transparency, accountability, and non-discrimination in AI systems.[3]
Further, in the EP’s negotiating position, the MEPs streamlined their attempts to boost citizens’ rights by allowing for the filing of complaints about AI systems and receiving explanations of decisions based on high-risk AI systems that significantly impact their fundamental rights. [4] Beyond this, AI systems with an unacceptable level of risk to people’s safety would therefore be prohibited, such as those used for social scoring. MEPs expanded the list also to include bans on intrusive and discriminatory uses of AI.[5]
The AI Act also emphasizes its interplay with the GDPR, establishing specific requirements for data usage in AI systems, including the principle of data minimization and the use of privacy-preserving techniques.[6]
Through setting high standards, the Act fosters a responsible approach to AI development and use, ensuring that technological advancements align with the values and rights within the Union. Paving the way and crafting hard laws on AI, the EU is in a position to influence global practices, establishing a benchmark for other countries to develop their own frameworks that prioritize ethical considerations and fundamental rights in the deployment of AI.
5. EU and Council of Europe’s systems brought to unity?
The Council of Europe’s upcoming international AI Convention is making considerable progress, and the European Commission (negotiating on the behalf of the EU) is committed to actively engaging with the Council of Europe in this process to work toward an agreement on binding common AI principles and norms, bridging the institutions in the context of AI.[7]
The Council of Europe’s Committee on Artificial Intelligence (CAI) is legislating on development, design, and application of AI based on the standards for human rights, democracy, and the rule of law. Expected to reach an agreement by November 2023 and to be of legally binding nature, the new treaty will be open to Council of Europe members, the EU and third countries who share the same values. [8]
While the AI Act is the first regulation on AI directly applicable in all 27 EU Member States, the future AI Convention will have to be implemented in the parties’ legal order through domestic legislation. If the EU joins the convention, the EU AI act becomes a mechanism through which the Council of Europe convention on AI can be integrated into the EU legal system as a whole.
6. Exploring the global race: implications and potential improvement to the overall AI governance system
While considering the next steps, it is important to note that the EU has reiterated its strong and continuous support for the Council of Europe’s work to promote fundamental rights, especially in the field of AI. The EU’s key objectives are to take a ‘risk-based and future-proof strategy that leads to proportionate, effective, and unambiguous AI guidelines, while ultimately developing a flexible framework that avoids overlaps and adds value to other relevant instruments and conventions while remaining compliant with the AI Act. To bring both the EU and Council of Europe’s pending legislations into unity, it is important to harmonize their stances within this tightrope of innovation and regulation.
The current stamina of regulators to delve into the realm of AI brings both opportunities and concerns, given the high risks that accompany AI technologies. If the EU institutions and the Council of Europe, as well as relevant policymakers and global actors, collaborate and partake in negotiations to ensure the guarantee of fundamental rights of their citizens, then upcoming regulations will bring further improvement to the overall AI governance system by harnessing the real potential and transformative power of AI for the advancement of society.
[1]European Parliament, AI Act: a step closer to the first rules on Artificial Intelligence, europarl.europa.eu, 11 May 2023.
[2] B. Kohn – P. Fritz-Ulli, AI regulation around the world, in taylorwessing.com, 9 May 2023.
[3] European Parliament, Artificial Intelligence, in europarl.europa.eu/thinktank, 23 March 2023.
[4] Heywood D., The EU’s AI Act heads towards final negotiations, in taylorwessing.com, 24 May 2023.
[5] Ibid.
[6] G. Gitchev, The Governance of the AI Act: your questions answered, in futurium.ec.europa.eu, 4 March 2022.
[7] Delegation of the European Union to the Council of Europe, EU participates actively in Council of Europe negotiations for the development of a new convention on artificial intelligence, in eeas.europa.eu, 30 January 2023.
[8] Ibid.
