A new era for regulation: The rise of Generative AI
Artificial Intelligence (AI) has been steadily shaping our world, but the recent rise of Generative AI (GenAI) marks a significant turning point. While regulators are accustomed to adapting to technological advancements in their industries, the unprecedented speed at which GenAI is driving change presents a new challenge. It's no longer a question of if AI will impact regulators and regulated industries, but how. The transformative potential of GenAI is accelerating the adoption of all flavours of Artificial Intelligence, requiring regulators to develop a unique position on not only how to regulate it, but also how to leverage its capabilities for a more effective approach to regulating.
To explore how this impacts the world of regulators, Deloitte recently partnered with the Institute of Regulation (IoR), bringing industry leaders together with over sixty regulators from various sectors to share insights, concerns, and potential solutions surrounding AI. We set out five key takeaways that emerged from these conversations.
1. Evolving regulatory landscape - a principle led approach:
The UK government has expressed a clear desire for a principle-led approach to AI regulation and whilst they have published a pro-innovation framework, there are no plans for AI-specific legislation and existing laws such as the Data Protection Act and the Equality Act will provide a foundation for addressing some AI-related challenges. Given the rapid pace of AI’s technological evolution, any principle-based framework must be agile and adaptable. For example, many UK companies operate internationally and may be subject to more prescriptive approaches to regulating AI, e.g. in the EU. Regulators may find it helpful to examine international frameworks (such as the EU AI Act) to understand the wider landscape that their regulated entities are operating within as part of developing their own guidance.
2. Balancing innovation and risk - keep a human in the loop:
There was a recurring theme, which is familiar to regulators, of the need to balance enabling AI innovation amongst companies operating in the UK whilst mitigating potential risks and harm to UK markets and society. Some regulators suggested a "permissively cautious" strategy, such as providing sandbox environments for organisations to explore tech development whilst also assessing and identifying risks that may come with the use of AI - this has proven to be a successful approach in sectors such as aviation. On the other hand, some regulators favoured a more precautionary stance in limiting the use of AI, due to the potential biases and unintended consequences of AI use. This tension was highlighted in discussions on GenAI "hallucination", with regulators expressing concerns about the trustworthiness and transparency of AI tools.
The need to balance the potential benefits of AI with robust oversight highlighted the need for frameworks such as Deloitte’s Trustworthy AITM framework and the need to retain a human in the loop.
3. Industry perspective - navigating uncertainty and seeking clarity:
Industry leaders expressed a need for greater regulatory clarity regarding the application of AI regulations to their specific industries and operations, this was supported by Deloitte’s ‘The State of Generative AI in the Enterprise’ report, which highlighted that 78% of leaders surveyed wanted more government regulation of AI. Perspectives on the desired level of regulation varied, with some advocating for more prescriptive guidelines, and others seeking flexibility to foster innovation. However, recognising that this is unlikely to be forthcoming in the shorter term, there is an increasing need for transparency and consistent communication between regulators and industry to support the collaborative approach needed to navigate the complexities of AI implementation. By engaging in regular, direct, and open dialogue, regulators can make a more informed decisions when designing policy and implementing rules to balance the need for innovation whilst managing for these risks.
4. Small investment, big impact:
Embracing AI is not limited to expensive and complex solutions and nor need it stall at the proof of concept stage. Regulators, particularly those with limited resources, are demonstrating the value of starting small with readily available, and often inexpensive, AI solutions. One smaller regulator shared their success in operationalising an AI solution to improve the effectiveness and efficiency of their day-to-day processes. This success story highlights a crucial point: regulators can achieve significant efficiency gains and process improvements by embracing practical AI solutions and adopting an agile, iterative approach – starting small, thinking big, and testing often.
5. GenAI - from everyday efficiency to future transformation:
While many regulators are already exploring GenAI, they are demonstrating a stronger preference for using it to enhance internal efficiencies - rather than deploying it in outward-facing applications involving the public or regulated entities. This is primarily driven by risk appetite. However, there is also increasing interest in more advanced solutions, such as Deloitte's NavigAite, that uses GenAI for a wide range of regulatory activities including accelerating regulatory investigations, improving customer experience and optimising budgets. Focus is further shifting to wider RegTech to support the way that regulation is delivered, with concepts such as Norm Engineering that uses GenAI to translate complex legal texts into practical specifications or executable models. These advancements point towards a future where GenAI can fundamentally transform how regulation is developed, consumed, and implemented.
While GenAI offers regulators significant opportunities to modernise their approach to developing and implementing regulation, securing the skills and capabilities (human or machine) alongside limited resources remains a challenge, demanding strategic planning and an innovation mindset.
Looking ahead:
The discussions emphasised the critical and unique role regulators play in navigating the expanding AI landscape in UK markets and beyond. Moving forward, regulators may benefit from the following key considerations:
- Work collaboratively in dialogue with industry to explore AI and codify this into guidance that is regularly refreshed to meet the evolving industry needs.
- Leverage existing frameworks and tools such as Trustworthy AITM to better understand and mitigate key risks.
- Improve collaboration and communication across regulators, government, and wider industry to overcome the challenges and maximise the opportunities.
- Identify low risk operational areas where AI, particularly GenAI, could quickly improve the efficiency of regulators to free up valuable resources.
- Embed more effective strategic planning and prioritisation alongside more robust business cases so organisations benefit in both the short term and longer term.
Concluding thoughts:
To effectively address the opportunities and challenges of AI, amplified by the rise of GenAI, the regulatory landscape must evolve. These takeaways, from our conversations with regulatory and industry leaders, highlight the pressing need for collaboration, innovation, and a commitment to accelerate its adoption whilst upholding ethical considerations. Creating a future where AI is harnessed responsibly for the benefit of all.
We’d like to thank everyone who brought their insight to our conversation on AI regulation and for allowing us to share these observations.
