AI View - August 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. The UK Government publishes its AI Opportunities Action Plan: terms of reference as a precursor to setting out its roadmap to an AI Bill

2. The EU delays the release of its technical standards for the AI Act

3. Japan moves towards passing AI legislation

4. The UAE introduces a Charter for the Development and Use of AI

5. The UK Department of Science, Technology and Innovation launches AI cybersecurity consultation

6. The US Department of Commerce announces new guidance to help AI developers evaluate and mitigate AI risks

The UK Government publishes its AI Opportunities Action Plan: terms of reference as a precursor to setting out its roadmap to an AI Bill

On 26 July 2024, the UK Government released its AI Opportunities Action Plan: terms of reference, which outlines its plan to develop a roadmap towards an AI Bill that will capture the opportunities of AI and harness its potential for economic growth.

The Action Plan will be developed with input from leading experts on AI from business and academia, and will consider how the UK can:

• Develop a scalable and globally competitive AI sector.

• Utilise AI to boost economic growth and productivity.

• Improve public services with AI.

• Transform citizen interactions with the government and increase AI adoption across the public sector and wider economy.

• Enhance the foundations for AI adoption through improvements in data, infrastructure, procurement, policy, and regulatory frameworks.

Read the full announcement here.

The EU delays the release of its technical standards for the AI Act

It has been reported in the press that the technical standards for the EU AI Act have been delayed. These standards aim to help businesses achieve technical compliance with the requirements of the AI Act, particularly for high-risk AI systems. Work has been underway to develop these standards, but they are now expected to be completed by the end of 2025, approximately eight months later than initially anticipated, according to the chair of the group developing the standards.

This means that the standards will be ready six months prior to when the majority of the AI Act's provisions enter into force, in August 2026. Concerns have been raised in the AI sector that the standards will not be ready in time for the Act's compliance deadlines, which would make meeting its obligations significantly more difficult.

Despite that, the chair of the AI standards body indicated that the standards would be ready on time and would be of a sufficient quality to ensure they provide the guidance that the industry needs to comply with the Act.

Japan moves towards passing AI legislation

It has been reported in the press that the Japanese government is moving towards passing specific legislation for AI, shifting from its initial preference for a risk-based "soft-law" approach to AI regulation.

The government held meetings with a newly formed panel of experts, including academics and AI industry leaders, to consider expanding existing AI guidelines, which are largely self-regulating. The meeting addressed the inadequacies of the current AI guidelines and "blind spots" within Japan's rapidly evolving AI sector, with a focus on Japan-specific legislation.

The panel plans to use the current guidelines as a foundation to develop measures based on the level of AI risk. A key consideration involves designing flexible legislation that can keep pace with rapid technological innovation to ensure any laws are in line with international standards such as the EU AI Act.  

The UAE introduces a Charter for the Development and Use of AI

On 30 July 2024, the UAE's Artificial Intelligence, Digital Economy and Remote Work Applications Office introduced the Charter for the Development and Use of Artificial Intelligence. It forms part of the UAE Strategy for AI, which aims to transform the UAE into a global hub for developing and adopting AI.

The charter emphasises creating a safe, private and trust-enhancing environment for AI applications, promoting societal benefits and inclusivity. It sets out principles for safe and fair AI development, aiming to strengthen human-machine relationships, enhance AI awareness, ensure equitable technological access and comply with relevant laws, marking a significant step towards enhancing the UAE's digital landscape and global AI partnerships.

Read the full announcement here.

The UK Department of Science, Technology and Innovation launches AI cybersecurity consultation

The UK Government recently opened a consultation process to establish a voluntary code of practice for AI cybersecurity, aiming to enhance the safety of AI technologies through robust measures, saying it is "crucial" that AI safety is underpinned by cybersecurity.

This initiative, announced by the Department for Science, Technology and Innovation, sought to gather input on the cybersecurity of AI with hopes that the development of a code of practice could set a new global standard.

The consultation is part of a broader strategy to secure and advance the UK's digital infrastructure, including AI and connected technologies, and was accompanied by a related consultation for software vendors.

Read more here.

The US Department of Commerce announces new AI guidance to help AI developers evaluate and mitigate AI risks

On 26 July 2024, as contemplated in President Biden's Executive Order on the Safe, Secure and Trustworthy Development of AI, the US Department of Commerce released:

• final and draft guidance documents aimed at mitigating risks associated with AI; and

• a new software package for testing AI system vulnerabilities to adversarial attacks.

The US Patent and Trademark Office also updated its guidance on patent subject matter eligibility for AI innovations. These efforts are designed to (a) inform developers about the unique risks of generative AI and (b) support further innovation while managing these risks.

Find out more here.