AI View - May 2024

Welcome to AI View, Simmons & Simmons’ fortnightly round-up of key AI legislative, regulatory, and policy updates from around the world. This edition brings you:

1. EU AI Act set to come into force in June / July of this year

2. France and China’s joint declaration on AI and global governance

3. Colorado publishes “first-in-the-nation” Artificial Intelligence Act

4. OECD Ministerial Council adopts revisions to its Principles on AI

5. The 9th UK-Singapore Financial Dialogue

EU AI Act set to come into force in June / July of this year

The EU AI Act is expected to be approved by the ministers at the Transport, Telecommunications and Energy Council meeting in Brussels on Tuesday 21 May 2024.

This meeting will set in motion the adoption of the Act, which is expected to come into force between the end of June and mid-July.

The current roadmap for the implementation phase will include the following steps:

1. Companies will have until January / February 2025 to stop using prohibited AI technologies and to ensure, to their best extent, a sufficient level of AI literacy of their staff;

2. from June / July 2025, General Purpose AI (GPAI) models, regulatory and penalty provisions will apply;

3. June / July 2026 will mark the end of the implementation period and the classification rights for high-risk AI systems will apply – at this date, Member States will (as a minimum) need to show that at least one AI regulatory sandbox (at a national level) is operational; and

4. In June / July 2027, the delayed application of requirements for high-risk AI systems will apply, and the grace period for non-compliance of existing GPAI models will end.

Read the proposed agenda (French language) here, and the EU Council article here.

France and China provide joint declaration on AI and global governance

On 6 May 2024, France and China concluded President Xi Jinping's state visit to France by issuing a joint statement on AI. The visit marked the 60th anniversary of bilateral relations between the two countries, reaffirming their commitment to international cooperation in the field of AI.

Recognising AI's key role in development and innovation, the nations agreed on the importance of promoting AI development and security and fostering appropriate international governance.

The Joint Statement sets out:

• a commitment to promoting safe, secure and reliable AI systems based on the principle of AI for the common good;
• an agreement to promote AI development and security;
• an agreement to contribute to strengthening cyber capacities of states;
• an acknowledgement of AI’s potential positive impact on the future of work; and
• a pledge to deepen discussions on global AI governance.

Whilst China has been discussing AI with EU states and other Western nations, this joint statement marks the first signed commitment to collaborate and cooperate with a major Western nation on policy and governance of AI.

Read the original press release (in French) here.

Colorado publishes “first-in-the-nation” Artificial Intelligence Act

On 8 May 2024, the Colorado legislature passed a bill on AI which includes a mandate for developers and deployers of high-risk AI systems to exercise reasonable care to prevent algorithmic discrimination.

If signed into law, Bill SB24-205 would become the first state-level legislation regulating the use of high-risk AI systems, and would come into effect on 1 February 2026. Following the EU AI Act, the Colorado AI Act takes a risk-based approach which applies to “developers” and “deployers” and sets out certain duties on deployers around: (i) the creation of risk management programmes, (ii) notification to the Attorney-General’s office and (iii) allowing a right to appeal adverse consequential decisions.

The Bill will also require businesses using AI systems that interact with consumers to disclose to the consumer that they are interacting with an AI system, with the Attorney-General being granted exclusive authority over the enforcement of the bill and the sanctions therein.

There is a carve-out available, allowing developers and deployers of AI systems to avoid enforcement action if they have taken affirmative action to implement the technology with a recognised risk management framework for AI systems, and have taken measures to monitor for any violations of the Bill.

Read more from the Colorado General Assembly website here.

OECD Ministerial Council adopts revisions to its Principles on AI

On 3 May 2024, the OECD Ministerial Council Meeting (MCM) approved the adoption of changes to certain AI definitions and concepts as part of its flagship OECD Principles on Artificial Intelligence. The Principles were adopted in 2019, and it was recognised that they required an uplift to remain fit-for-purpose in light of the proliferation of the use of general purpose and generative AI. The amendments directly address AI-related challenges involving privacy, safety, IP rights and more.

The updated Principles address:

1. mitigating safety concerns, to protect against the risk of AI systems causing undue harm or exhibiting undesired behaviour;

2. reducing the risk / impact of mis- and disinformation and safeguarding information integrity in the context of generative AI;

3. emphasising responsible business conduct throughout the AI system lifecycle;

4. refining and defining the need for jurisdictions to work together to promote interoperable governance and policy environments for AI;

5. clarifying expectations around transparency and responsible disclosure in relation to the use of AI systems; and

6. explicitly referencing and incorporating environmental sustainability and related language.

Read the OECD update here, and the Principles here.

The 9th UK-Singapore Financial Dialogue

On 8 May 2024, Singapore and the United Kingdom held the 9th UK-Singapore Financial Dialogue.

Both countries engaged in discussions concerning risks to financial stability and cyber security, along with supervisory issues related to the growing application of AI in the financial sector.

There was accord between the nations that international standards play a crucial role in promoting innovation whilst controlling risks, and there was discussion considering potential areas for collaboration via international platforms such as the International Organization of Securities Commissions (IOSCO) and the Federation of Small Businesses (FSB).

As part of this dialogue, an industry-led UK-Singapore business roundtable on generative AI was held on 9 May, with participation by industry stakeholders from both countries. The roundtable discussed the challenges to regulatory compliance brought about by generative AI, and identified strategies to harness the benefits of AI whilst mitigating risks.

These discussions underscore the increasing focus on AI within the financial sector, and demonstrate a shared dedication between both nations to find collaboration opportunities in this context.

Read the HM Treasury update here.