Quick Buzz Feed
This article discusses how administrative law should adapt to the rise of artificial intelligence, specifically focusing on the United Kingdom's evolving approach. It examines the foundational tension between traditional legal frameworks and the governance of AI-assisted decisions, outlining the UK's search for a new legislative framework. The questions animating this symposium—how administrative law should adapt to the rise of artificial intelligence—are hardly confined to the United States. The United Kingdom, like many other jurisdictions, is grappling with the same foundational tension.
The United Kingdom, much like the United States, is actively addressing the complexities of integrating artificial intelligence into administrative law. A central inquiry revolves around whether existing administrative law frameworks are adequate to govern decisions made or supported by AI, or if new, purpose-built legislation is essential to address the unique risks and innovative opportunities that AI presents in public administration. The UK's stance on this issue is rapidly evolving, with recent developments indicating a strong movement towards the serious consideration and potential implementation of a custom legislative framework for AI governance.
The UK government has demonstrated a remarkable pace in embedding AI across various sectors of public administration, driven by aspirations for improved outcomes, reduced costs, and increased operational scale. While implementation has been somewhat inconsistent, the use of AI tools is becoming widespread. Judges are now openly utilizing AI, and numerous large-scale administrative deployments are evident. Over a hundred examples are publicly listed on a government-led transparency register. Concrete applications include AI tools for detecting anomalous Value Added Tax (VAT) returns to identify potential fraud, the widespread adoption of chatbots by governmental departments like the Department of Work and Pensions and the Driver and Vehicle Licensing Agency for triaging and prioritizing inquiries, and the use of AI-based risk assessment models, such as the Universal Credit Advances Model, to combat welfare fraud.
Following Brexit, the UK initially adopted a 'pro-innovation' strategy towards AI regulation, emphasizing experimentation over upfront legislative control. While traditional administrative law principles—including common law judicial review, statutory procurement rules, data protection law, and equality legislation—are theoretically applicable, their practical effectiveness in governing AI remains limited. Very few disputes involving AI in administration have culminated in full judicial decisions, and existing enforcement mechanisms struggle to achieve significant impact. Despite the development of 'soft law' guidance by public bodies to interpret existing norms for AI implementation, there is profound uncertainty regarding the capacity of current administrative law frameworks to effectively manage the use of AI within public administration.
The UK's previous 'wait-and-see' approach to AI regulation in public administration, though intended to foster innovation, has increasingly revealed significant inherent risks. The absence of robust and enforceable legal regulation heightens the vulnerability of individuals to failures arising from both individual and systemic administrative errors in AI-driven processes. Furthermore, this approach risks creating a fragmented and inconsistent regulatory landscape, with overlapping but often limited jurisdictional powers among different regulators and sectors. Such legal uncertainty generates considerable challenges for administrative bodies, making it difficult to confidently apply established law to rapidly evolving AI systems. A single adverse court ruling could have far-reaching implications for government operations increasingly reliant on these technologies, as demonstrated by international examples such as Australia's 'Robodebt' scheme and the 'SyRI' system in the Netherlands, both of which severely damaged government credibility following high-profile failures.
In response to the growing recognition of these accumulating risks, the UK is now visibly shifting away from its prior 'wait and see' attitude towards a more intentional regulatory strategy. The policy discourse has increasingly focused on the imperative of developing a distinct UK framework for governing AI's use in administrative actions. This significant effort is notably being spearheaded by the Law Commission, an independent statutory body responsible for reviewing and recommending legal reforms for England and Wales. In its recently unveiled Fourteenth Programme of Law Reform, the Commission explicitly underscored the critical importance of this moment, describing the development of a coherent legal framework for automated decision-making as 'the most significant current challenge in public law,' highlighting the potential harms to the public and public confidence should these systems fail. The Commission has committed to exploring various legislative reform options, including the possibility of an overarching statutory framework or more targeted reforms within specific administrative sectors. This commitment, typically sponsored by government departments, signals a profound change from the UK's previous hands-off approach to a direct engagement with legislative intervention.
While the Law Commission's process for legislative reform will extend over several years, UK institutions, including courts and public bodies, are tasked with the immediate and complex challenge of applying existing, often outdated, legal frameworks to novel AI technologies that fundamentally challenge their underlying assumptions. This situation underscores the critical importance of ongoing dialogue and shared learning between the administrative law systems of the US, UK, and other jurisdictions globally. As the modern administrative state continues to evolve with the integration of AI, collaborative efforts and the exchange of insights will be invaluable in navigating these unprecedented legal and governance challenges. The conversation between nations is more vital now than ever before, shaping how AI will be governed in the realm of public administration.