UK AI regulation and what this means for UK innovation and growth

The reason behind this announcement is closely connected to the EU AI Act, which identifies OpenAI as high-risk and stifles innovation.

Unlike the EU AI Act, which carries legal weight, the UK Department for Science, Innovation, and Technology (DSIT) has recently released a new white paper titled ‘AI regulation: a pro-innovation approach.’ This paper promises to only legalise AI when necessary, presenting a different approach to regulating AI technologies.

Some key differences also include the UK's broader definition of AI, allowing for future-proofing, and a context-specific approach that evaluates outcomes before implementing regulations. This approach encourages responsible innovation and sets it apart from the EU AI Act, which imposes restrictive rules that can hinder normal operations and even push AI enterprises like OpenAI to potentially exit the market.

In this article, Dr Leslie Kanthan, CEO and Founder of TurinTech, delves into the UK AI regulations and explores the key concepts outlined in the white paper. He also aims to uncover the implications of these new rules for businesses and consumers.

An overview of the White Paper

AI risks + current regulatory environment = need for new regulation

AI, like a double-edged sword, holds enormous potential across diverse fields such as medicine, science. However, we must also acknowledge its capacity for harm and negative implications due to insufficient control, biased training data, and other factors. These risks range from threats to human rights and safety to privacy breaches and societal well-being. To ensure the responsible and beneficial use of AI, proactive measures are crucial.

But that’s not all. The current regulatory environment is another factor provoking the need for changes. Why? First of all, existing laws lack specific coverage of AI products, which leaves consumers vulnerable to AI risks. And secondly, despite multiple regulatory bodies supervising the AI industry, there is a lack of consistent coordination among them.

Government objectives

The UK government is committed to promoting growth and prosperity by fostering responsible innovation and reducing regulatory ambiguity. In addition, they prioritise improving public trust by actively addressing and resolving risks associated with AI. By accomplishing these objectives, they aim to consolidate the UK's position as a global leader in AI, fulfilling their ambition on a broader scale.

Government approaches

• Pro-innovation: AI development is not undermined but regulated to mitigate potential risks.
• Proportionate: The focus is on regulating the use of AI, considering context and outcomes, rather than targeting the entire technology industry. This approach allows tech companies to prioritise product innovation without excessive compliance burdens.
• Trustworthy: Addressing public concerns about AI risks to build trust and encourage the adoption of AI technologies.
• Adaptable: Enabling the government to keep pace with the evolving opportunities and risks associated with AI technologies.
• Clear: Ensuring clarity and simplicity throughout the AI lifecycle for businesses using AI by providing clear rules, specifying the enforcement authority, and offering compliance guidance.
• Collaborative: Promoting cross-sector collaboration among government, regulators, and industry to foster AI innovation and ensure that diverse opinions are valued.

Five principles of AI regulation

• Safety, security, and robustness: AI systems should operate with safety, security, and robustness in mind.
• Transparency and explainability: AI businesses should be able to explain the purpose, workings, deployment plans, and decision-making processes of their AI systems.
• Fairness: Individuals should not be subject to discrimination, and their legal rights should not be undermined.
• Accountability and governance: Businesses should take action to appoint a responsible person to ensure effective supervision of AI systems.
• Contestability and redress: AI systems should be capable of challenging and correcting harmful AI outcomes.

Impacts on foundation models and LLMs

Before exploring the impacts of this regulation on consumers and business leaders, let’s briefly examine its implications for two specific and impactful types of AI: Foundation models and LLMs (Large Language Models), as outlined by the DSIT in the white paper.

Foundation Models, as their names imply, serve as the foundation or starting point for various AI tasks. They devour vast amounts of data to help them learn patterns and acquire a general understanding of text, images, or other types of information. However, comprehending their inner workings, capabilities, and potential risks can be complex. Risks such as privacy violations and biases may arise. To ensure accountability, DSIT acknowledges the need for comprehensive oversight throughout the lifespan of a foundation model. The new rules propose centralised control through central functions, employing technical standards and assurance techniques for effective compliance and evaluation. On the other hand, the white paper cautions against hasty regulatory actions targeting foundation models and LLMs, recognising the potential adverse impact on their beneficial applications. For LLMs, guidance on transparency measures is suggested, with ongoing monitoring and evaluation by the government.

The impact of UK AI Regulation on AI businesses and consumers

What does the regulation plan mean for you, whether you are a business leader looking to adopt AI within your organisation or an end consumer of AI products? The following section aims to address your concerns by examining the impact assessment presented in the white paper released by DSIT.

Benefits:

• Encourages responsible innovation: AI regulation empowers business leaders to actively contribute to the development of innovative solutions while ensuring ethical and safe practices. By adhering to these standards, you have the opportunity to drive responsible and sustainable innovation that benefits the public and reduces potential risks.
• Increased clarity and trust: AI regulation benefits both businesses and consumers, fostering clarity and transparency. As a business leader, you can navigate the AI landscape confidently, knowing the rules and expectations. As a consumer, you can trust AI products and services, ensuring privacy and safety. This empowerment strengthens the bond between businesses and consumers, fostering confidence, loyalty, and growth.
• Data sharing: When you have gained public trust as a business leader, you can create an environment where consumers willingly share their personal data with your business. This data becomes a valuable resource for driving innovation and improving product performance, benefiting both you and your consumers.
• Enhanced safety and quality: Government regulations ensure that AI systems meet specific requirements, resulting in higher quality, safer AI systems with reduced biases. As a consumer, you will find the AI technologies you use to be more accurate, reliable, and applicable.

Costs:

• Playing by the rules can be costly: Meeting regulatory requirements comes with expenses for businesses. This includes training employees and making changes. For small businesses, these costs can be relatively higher because they represent a fixed amount, taking up a larger portion of the company's overall expenses.
• Potential price increases: If businesses pass on the costs of regulatory compliance to consumers, it may lead to a price increase for AI products and services. For consumers, this means you may need to consider the affordability and value of these offerings, while business leaders should consider potential negative effects, such as a decrease in sales volume.
• Limitations and prohibited outcomes: Some AI systems or outcomes, even if not harmful, may be restricted due to regulatory compliance. This may limit the effectiveness of AI systems in helping consumers achieve their goals.
• Missed opportunities: An increase in costs, prices, and AI prohibition can lead to financial losses for businesses. This hinders their ability to invest in improving AI products and services. Consequently, consumers miss out on enhanced AI offerings that could have improved their experiences and catered to their interests.

Future outlook

The regulatory journey of the UK is just getting started, seeking feedback and better solutions from businesses, regulators, and individuals. The aim is to ensure the implementation of effective regulation while minimising adverse impacts and avoiding hindrances to innovation.

Following the consultation period, the government aims to consider the feedback received from various stakeholders and respond by publishing strategic plans for AI regulation. In the coming year, regulators will be expected to develop guidance documents that aim to fulfil the five principles outlined in the white paper.

In addition to the white paper, the Prime Minister announced on 7 June 2023 the hosting of the first global summit on AI safety in September. This gives another opportunity to actively contribute to the development and implementation of AI regulations. Make sure to stay engaged and well-informed about the ongoing discussions regarding AI regulation.

References:

https://www.jdsupra.com/legalnews/uk-government-proposes-new-ai-1590962/

​https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach/white-paper

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1147045/uk_ai_regulation_impact_assessment.pdf

https://datainnovation.org/2023/04/an-overview-of-the-uks-new-approach-to-ai/#:~:text=Under%20the%20new%20approach%20to,by%20or%20for%20the%20government.

https://time.com/6282325/sam-altman-openai-eu/