Knowledge

Designed to help organisations make decisions about access to sensitive data: safe projects, safe people, safe data, safe settings, safe outputs. Developed by the ONS and now the backbone of Secure Data Environments and Trusted Research Environments across the NHS.

The clinical safety standard that applies to any health IT system whose malfunction or misuse could harm a patient. Covers clinical safety cases, hazard logs, and the requirement for a designated Clinical Safety Officer. Applies to most AI tools entering clinical pathways.

The deploying organisation's counterpart to DCB0129. NHS trusts accepting health IT systems must conduct their own clinical safety assessment in their own context — they cannot simply accept the vendor's safety case. Most organisations do not know they hold this obligation.

The global standard for enterprise architecture. Its four domains — business, data, application, technology — remain relevant for AI deployments, though most organisations skip straight to application and technology. That is precisely where AI strategies fail.

The most extraordinary manufacturing feat in human history: firing lasers at tin beads to generate plasma hotter than the sun, producing chips with features measured in nanometres. Essential context for anyone thinking seriously about AI's physical cost.

Anthropic's initiative exploring AI safety, transparency, and the long-term trajectory of frontier AI development. Relevant to anyone thinking about responsible AI adoption and what 'AI for good' actually requires in practice.

The global standard for defining digital and technology skills. The Future Roles section uses SFIA levels (3–7) and skill codes to ground AI-augmented roles in an established professional framework — connecting emerging human-AI capabilities to a language organisations already use for workforce planning.

NHS England's guidance on information governance for AI — covering data protection obligations, transparency requirements, and accountability frameworks specific to health and care settings. The bit most AI projects in the NHS skip until legal asks why they skipped it. Essential reading before any AI system touches patient data.

The UK government's practical guidance for public sector organisations adopting AI — covering how to identify opportunities, assess readiness, procure responsibly, and govern deployment. Written for decision-makers rather than technologists. Useful framing for NHS and government bodies navigating AI adoption without a clear internal playbook.

The world's first comprehensive legal framework for artificial intelligence — risk-tiered, with outright prohibitions at the top and lighter obligations for minimal-risk systems. Healthcare AI sits firmly in the high-risk category. If you are deploying AI in a clinical or regulated setting, this is not optional reading: it is the compliance baseline that will shape procurement, governance, and accountability across the NHS and beyond.

A structured lens for reading the environment an organisation operates in: Political, Economic, Sociological, Technological, Legal, Environmental. Simple in concept, deceptively powerful in practice. Most strategy work skips it — which is why most strategies fail to anticipate the conditions they land in. Essential groundwork before any AI adoption plan that expects to survive contact with reality.

The physical substrate that AI conversations consistently gloss over. 67.7GW of global data centre capacity — up 36% in two years — now accounts for 2% of the world's electricity. The US alone consumes 43% of that footprint. Germany runs 9.5% of its national grid through data centres; the UK runs 5.8%. These are not technology statistics — they are energy infrastructure decisions with thirty-year consequences. Essential grounding for anyone thinking seriously about AI sustainability, national digital strategy, or the real carbon cost of cloud-native transformation.

Live visibility into how green Great Britain's electricity grid is right now. NESO — the National Energy System Operator — publishes real-time carbon intensity, generation mix, and grid balancing actions in one place. Useful context for anyone thinking about AI infrastructure, data centre energy sourcing, or what net zero looks like in operational practice. The gap between the ambition and the live numbers is instructive.