I’m reading through the 10 Downing Street AI Opportunities Action Plan. At first glance it’s sharp and comprehensive. One little recommendation alone — which is really only adjacent to AI itself — could have enormous transformational effects on the efficiency of G7-scale governance in the 21st century.
44. Use digital government infrastructure to create new opportunities for innovators. For example, an approach akin to Jeff Bezos’s API mandate at Amazon could be adopted. This required all teams’ data and functionality to be exposed through APIs (Application Programme Interfaces). All standard documentation interactions, like compliance or planning, could be done through APIs, to which companies could connect their own tools. Similarly, mandating e-Invoices from government suppliers could automate billing, speed up payments and reduce fraud.
Now, I have my hesitations about much of what Amazon and Jeff Bezos represent, but the depth of their innovation and its impact on the global economy, and on people, is undeniable. I’m reminded of an incredible deep dive on Amazon’s 2002 API edict and how it was fundamental to avoiding death by bureaucratic overgrowth.
The author, who has insight into the inner workings of Amazon, said this:
In 2002, Jeff Bezos had the most important insight he would ever have: …the limiting reagent for Amazon’s growth would not be its website traffic, or its ability to fulfil orders, or the number of SKUs available to sell; it would be its own bureaucracy.
That is to say, Amazon scaled to the size of some nation state governments precisely because it designed a solution to exploding complexity into its operating system.
For the most part, governments didn’t have this opportunity — they had ongoing commitments to uphold while the Internet developed, and need to be much more risk averse. Now they’re moving through the challenging, but critical, task of reforming themselves for the digital age — swapping the engine out while the plane is flying. The AI Opportunities Action Plan suggests to me that Sir Keir Starmer and Peter Kyle understand the importance of swiftly and gracefully executing this transformation, both for the well-being of Britons and for the economic competitiveness of the UK in the 21st century and beyond.
Why is Amazon’s internal architecture relevant to AI?
It’s a deep topic, but as a general principle, the more predictable, well-structured, and well-documented digital services are — what data they accept, how they process it, and what data they return — the easier it is to build reliable software systems. One of the most incredible things about language models is that they can understand unstructured data — contemporary AI systems are overcoming this limitation computers have faced for decades.
Even with this advancement, though, the fact remains: predictable, stable data services are key to building software systems. Moving the UK government’s digital services towards a system architecture more like Amazon’s internal APIs could improve speed, efficiency and extensibility, providing “a solution to the bureaucratic complexity that would have otherwise caused it to implode. Instead of being a bureaucratic liability, Amazon’s sprawl would become a massive surface area of customer contact.”
Interestingly, since the internal workings of API services don’t really matter to external consumers, this design still leaves a lot of room for specialization and innovation.
All this said, I recognize there are complexities that lie outside of my understanding.
This is a wildly incomplete list of questions and challenges the UK government will need to address.
The power of Amazon’s approach is in its ability to tame complexity through well-defined, universally accessible interfaces. While governments aren’t startups, and can’t take commensurate risks, they can learn from successes in the private sector. I’m excited by the UK’s direction on AI and my chance to play a small role in it.