AI View: October 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. UK Government launches new Regulatory Innovation Office
2. G7 nations discuss competition aspects in the development of AI
3. Polish Ministry of Digital Affairs opens consultation on prohibited AI under the EU AI Act
4. Japan releases guidance on evaluating AI safety and red teaming AI systems
5. Japan opens consultation on generative AI market dynamics
6. Australia opens review of AI and consumer law

UK Government launches new Regulatory Innovation Office

On 8 October 2024, as signalled in the Labour Party’s election manifesto, the UK’s Department for Science, Innovation and Technology announced the creation of the new Regulatory Innovation Office (RIO). The RIO has been established in order to accelerate public access to and regulatory clearance for new technologies.

The RIO will work alongside various regulatory bodies to update regulations, speed up approvals and ensure a cohesive regulatory approach, thereby enabling faster and safer deployment of innovations. This effort aligns with the Government's broader strategy to stimulate economic growth and positions the UK as a leading destination for investment in emerging technologies.

Its mission will initially focus on supporting the growth of four “fast growing areas of technology making a difference to people’s lives”, including engineering, space, AI and digital in healthcare and connected and autonomous technology. As the office evolves, the scope of its mission will broaden to include additional technologies and sectors.

Read the announcement here.

G7 nations discuss competition aspects in the development of AI

At the culmination of the two-day G7 competition summit in Rome on 4 October 2024, representatives from the G7 member states and the European Commission issued a joint statement addressing competition concerns related to AI.

The statement highlights that despite the enormous potential of AI, the technology also poses significant risks to competition and associated areas such as IP, consumer protection and data protection, that should not be ignored.

The statement also outlines a set of guiding principles to ensure society can reap the benefits of AI whilst mitigating the associated competition and data protection risks, which include:

• Human innovation and copyright: The risk of AI, in particular generative AI systems, curtailing the earning power of artists and content creators is high, while the nature of the AI content creation market means that creators often operate in a monopsony in which large AI companies hold the power to create and shape the market;
• Consumer protection: AI-generated outputs have the power to distort consumers’ decision-making processes via incorrect or misleading training data; and
• Privacy and data protection: The need for large data sets via which to develop and train AI systems can lead to the misuse of consumers’ and users’ data. Such data must be handled and safeguarded in compliance with existing privacy rules and laws.

Read the statement here.

Polish Ministry of Digital Affairs opens consultation on prohibited AI under the EU AI Act

Following the implementation of the EU AI Act (the Act) at the start of August 2024, the Polish Ministry of Digital Affairs has commenced a consultation regarding the provisions of Article 5 of the Act, which set out the prohibited AI practices (provisions which will come into force from February 2025).

The consultation aims to gather comments, suggestions and recommendations from interested parties on the application of these Article 5 provisions, including the collation of examples of AI systems which may be caught under the designation of ‘prohibited’ AI systems.

The consultation is due to close on 31 December 2024.

Read the announcement (in Polish) here.

Japan releases guidance on evaluating AI safety and red teaming AI systems

The Japanese AI Safety Institute has released two sets of non-binding guidance, on evaluating AI safety and red teaming AI systems respectively.

The first guide, on evaluating the safety of AI systems, sets out 6 principles to consider when conducting safety evaluations. These are:
1. Human centricity: Human rights guaranteed by the Japanese Constitution or granted internationally should not be violated;

2. Safety: During the development or use of an AI system, any harm or damage to human health or property should be avoided;

3. Fairness: Efforts should be made to eliminate unfair and harmful bias and discrimination against specific individuals or groups based on race, gender, national origin, age, politics and/or religion;

4. Privacy: The privacy of users should be respected and protected in accordance with privacy and data protection rules and laws;

5. Security: Action should be taken to prevent the use or deployment of AI from being (un)intentionally altered or misused by ‘bad actors’ and rogue organisations; and

6. Transparency: Based on the social context in which the AI system is used, information should be provided to users, to the extent possible, to ensure they know they are using or interacting with an AI system.

The guide then provides 10 evaluation mechanisms for upholding these principles, focussing on controlling toxic outputs, ensuring data quality and repeated evaluation of the development of AI systems.

The second guide, regarding red teaming AI systems, provides examples of attack methods, stresses the importance of structuring ‘red teams’ and emphasises conducting red teaming exercises throughout the lifecycle of the AI. Key takeaways from the guidance are:

• the inclusion of training on key areas such as misinformation, disinformation and manipulation;
• introducing plausibility in training scenarios; and
• the need for training to be delivered on several different types of attacks.

Read the guidelines on safety evaluation of AI here and on red teaming AI systems here.

Japan opens consultation on generative AI market dynamics

On 2 October 2024, the Japan Fair Trade Commission opened a research and information request consultation regarding the rapid advancement of AI technology and its impact on market growth.

The consultation will feed into the publication of the recently released discussion paper titled ‘Generative AI and Competition’ which looks at both domestic and international generative AI markets and the proper integration of these technologies into economy and society.

Stakeholders have been invited to provide comments by 22 November 2024.

Read the announcement here.

Australia opens review of AI and consumer law

The Australian Department of the Treasury has opened a consultation on whether the existing technology-neutral consumer laws remain fit for purpose and are able to protect consumers who make use of AI systems.

This is considered in particular in terms of the suitability of the Australian Consumer Law (Schedule 2 of the Competition and Consumer Act 2010) to:

• protect AI consumers; and
• facilitate the safe and responsible deployment of AI by businesses.

The consultation has invited stakeholders to consider the existing principles, the existing remedies available and the mechanisms for allocating liability among manufacturers and suppliers. The responses will then feed into an ongoing review of Australia’s consumer protection framework, and more generally to clarify and strengthen existing laws addressing AI-related risks and harms.

Responses may be submitted until 12 November 2024.

Read more here.