The international regulatory landscape

Although artificial intelligence has been the subject of much discussion in recent months, no country has yet adopted comprehensive legislation governing AI. So far, broader-reaching legislative proposals are still under debate, and existing regulation is limited to certain sub-fields of artificial intelligence.

Even existing regulations are subject to debate and amendment, leaving the global regulatory landscape in its nascent stages. Certain regulatory trends are beginning to emerge, but there is still little clarity as to what form AI regulation will eventually take.

In this article, we take a brief look at some of the main legislative proposals on artificial intelligence around the world.

European Union

The European Union has a robust, wide-reaching proposal for the regulation of artificial intelligence, the Artificial Intelligence Act (AI Act), which has been under debate for some years. The first version of the proposed legislation was presented by the European Commission in April 2021. Consolidated proposed amendments were presented by the European Council in December 2022, and the European Parliament made other suggestions in June 2023.

The three institutions (the European Commission, Council and Parliament) are now discussing the final version of text. The expectation is that a consensus may be reached by the end of 2023 and the text finalized at the beginning of 2024, with the AI Act coming into force at the end of 2025 or the beginning of 2026.

The AI Act will apply to all areas where AI applications are used, creating global obligations for providers, users, importers and distributors of AI systems. The AI Act takes a risk-based approach to regulation, meaning that systems that present unacceptable risks, for example, will be banned from the outset, while those that present a high level of risk will be subject to additional obligations.

China

On 15 August 2023, the Interim Measures for the Management of Generative Artificial Intelligence Services were issued jointly by Cybernetic Administration of China and other Chinese regulatory authorities.

The Measures impose obligations on AI developers and users in China, including:

(i) security assessments of algorithms employed by generative AI service providers, 

(ii) labeling of AI-generated content so that it can easily be distinguished from other content; 

(iii) real-name verification of users, and 

(iv) implementation of anti-addiction measures and guidance for lawful use of generative AI.

Broader regulations on artificial intelligence, covering more than generative AI, are expected in the near future.

United Kingdom

Taking a very different stance from the European Union, the United Kingdom has adopted a self-proclaimed "pro-innovation" approach to the regulation of artificial intelligence. In March 2023, the UK government published a white paper on the subject, focusing on the idea of flexible, context-specific regulation to avoid creating obstacles to innovation.

The white paper presents an approach to regulation of AI based on five over-arching principles:

(i) safety, security and robustness, 

(ii) appropriate transparency and explainability, 

(iii) fairness; 

(iv) accountability and governance, and 

(v) contestability and redress.

The proposed model has encountered criticism, however. In August 2023, the House of Commons' Science, Innovation and Technology Committee published a report on the governance of artificial intelligence, indicating that the approach proposed by the government may position the UK unfavorably compared to other jurisdictions, which are advancing rapidly on AI regulation and may create standards which will then become international references.

United States of America

Like the UK, so far the United States has only a government document entitled White House Office of Science and Technology Policy, issued in October 2022, which sets out general guidelines and principles for the creation of an AI bill of rights.

Some months later, in May 2023, the White House made a series of announcements on safe, secure and responsible development of artificial intelligence, and obtained informal commitments from leading AI companies in the US.

To date there is no proposed federal legislation on artificial intelligence, although some states and municipalities have issued local rules on various aspects of AI use. For instance, the City of New York has established an ordinance regulating employer use of automated employment decision tools in hiring and promotions.

These are just a few of the jurisdictions that have been debating the regulation of artificial intelligence. The differences and similarities in their approaches underscores the importance of keeping a close watch on the development of AI regulation around the world.