The government must learn from years of experience introducing digital transformation if the UK is to make the most of the opportunities AI offers to the public sectorThe United Kingdom is currently in the throes of a transformative period, shaping the future of the public sector’s artificial intelligence (AI) capabilities for generations to come. The visible aspects of this transformation, such as the rollouts of Copilot, the establishment of foundation model partnerships, and the signing of high-profile contracts, are already well underway. However, the less conspicuous aspects of this transformation, such as the development of procurement capabilities, the cultivation of a diverse supply base, and the establishment of shared standards, are equally crucial. These elements will determine our ability to adapt as AI technology continues to evolve.
We have made significant strides in the first half of this task. However, the second half is what will determine whether the initial efforts bear fruit. This is what we refer to as the “silent lock-in” trap. It is the gradual accumulation of AI capabilities on top of infrastructure, management practices, and governance approaches that are individually defined, poorly coordinated, and mismatched to the rapid pace of technological change.
Despite the diligent efforts of individuals and teams to procure and experiment with emerging AI capabilities, the pieces are not fitting together as they should. So, how can we learn from the experiences of the past decade’s digital transformation to accelerate the UK’s adoption of AI? This is the central theme of the book, “Making AI work for Britain,” published by London Publishing Partnership and available for download at FutureOfAI.uk under an open-access license.
The book, based on several years of research into the UK’s AI strategy and ecosystem, and over a decade of experience working with the UK government on digital transformation, proposes a simple strategy for AI success – consolidate demand, diversify supply. The following is a summary of three key recommendations from the book: Build buyers who can push back; pool demand that is already shared; and keep the supply side plural.
The challenge of creating smart buyers is easy to articulate but difficult to solve. AI suppliers, especially the larger ones, often make claims that require significant technical expertise to evaluate. These claims pertain to aspects such as training data provenance, model behaviour under distribution shift, security properties of the fine-tuning pipeline, and interoperability with alternative providers.
Traditional procurement functions were designed to assess claims like, “this system meets this specification” and “this supplier has these references.” They were not designed to assess claims like, “this model will remain useful as underlying capabilities change.”
The smart-buyer model does not necessitate building deep AI expertise in every department. Instead, it requires each organisation that spends significantly on AI to have a small core of people who can sit opposite a vendor and understand what they are looking at. This core needs three things – technical understanding, commercial acumen, and the ability to influence the wider organisation.
In conclusion, the UK’s journey towards shaping the public sector’s AI capabilities is a complex one, fraught with challenges and opportunities. The success of this journey will depend on our ability to learn from past experiences, adapt to changing technologies, and build a robust procurement capability. The silent lock-in trap can be avoided by consolidating demand, diversifying supply, and building smart buyers who can push back. The future of AI in the UK’s public sector hinges on these critical factors.
