Goodbye Anthropic Road: Britain’s finally decided where its AI future lies
For a country that has ploughed its own furrow in the modern, global economy to become the world’s largest net exporter of financial services, Britain has become oddly dependent as a net importer of technologies developed elsewhere. That uncomfortable reality sits at the heart of the government's new Financial Services AI Adoption Plan, published on 14 July 2026 (“the Plan”). The document, developed by the sector's AI Champions, Harriet Rees and Dr Rohit Dhawan, paves a roadmap for accelerating AI adoption across financial services. It also, perhaps unintentionally, serves as a reminder of how much of the UK's digital future depends on decisions taken elsewhere.
The Plan is ambitious. It calls for greater regulatory clarity, a review of how AI-generated financial advice sits within the regulatory perimeter, stronger operational resilience measures, investment in skills and talent, and preparations for a future in which autonomous agents make payments on consumers' behalf. The underlying premise is simple: as the technology leaves the sandbox, Britain’s financial services firms need to be ready to scale their pilot programmes and capture the innovation, productivity, and growth benefits of AI.
Yet it was events on the very day of publication that illustrated the challenge most vividly. Reuters reported that many major British banks remain unable to access Anthropic's advanced AI model, Mythos, due to restrictions imposed by the United States. Access has reportedly been granted to a select group that includes American institutions, while large UK banks continue to wait. Rees described the situation as a “wake-up call”. This may prove to be an understatement.
This is not merely a question of commercial inconvenience. The Plan repeatedly highlights the growing dependence of financial institutions on a small number of cloud and AI providers. This is a harbinger of concentration, resilience, and security risks. The government's answer steers clear of protectionism. Instead, it calls for greater diversity of supply, more domestic capability, and a clearer vision of what “AI sovereignty” should mean in practice.
The timing is notable. The Mills Review, commissioned by the FCA and published on 6 July 2026 – in the week before the Plan’s release – envisaged a financial services sector that becomes increasingly “AI-enabled, continuous and delegated” by 2030, with AI embedded across firms, customer journeys, and market infrastructure. While the review concluded that existing regulation remains broadly fit for purpose, it also stated that governance, accountability, and consumer protection frameworks will become ever more important as AI systems become more autonomous.
Taken together, these developments suggest that the UK's AI debate is shifting. The question is no longer whether financial services will adopt AI; that seems inevitable. The more interesting question is whether Britain will remain primarily a consumer of frontier AI technologies created elsewhere, or whether it can build enough domestic capability to shape its own future.
For firms, that debate is no longer theoretical. It will influence technology strategy, operational resilience, outsourcing arrangements, talent acquisition, and regulatory engagement for years to come. Those seeking to understand where regulators may be heading should also consider CMS's analysis of the Mills Review.
The Plan is, at heart, a document about accelerating AI adoption. Read more closely, however, and it is also a document about loosening UK financial services firms from their dependence on non-sovereign AI tools. To remain a leader in this shifting landscape, it is time for Britain to go back to its plough.
