The introduction of machines to human decision-making

Historically decision-making has been a human endeavour. People made decisions, sometimes good ones, sometimes not. When decisions are very important, we take care to ensure they are made with a full understanding of the context and consequences and that they are well-framed, evidenced and unbiased. This is the basis of well tested decision-making doctrine in the military, business case processes in industry, patient pathways in health and safety protocols in power stations. 

As the world digitalises, the volume of available data explodes, and AI is adopted - intentionally and accidentally - into all aspects of life, the assumption that a human delivers all parts of a decision-making process is flawed. Increasingly machines are becoming part of the process, to pre-process data, develop and assess options, provide decision support or even take and effect decisions autonomously. This is because machines increasingly do many things more effectively, more efficiently and faster than humans. Machines are already part of today's decision-making so we must learn and equip ourselves for this, optimising the way humans and machines combine to inform, take and explain decisions.

Nowhere is it more important to get this right than in 'high stakes' decision making where the consequences of getting it wrong can be catastrophic  - serious loss of money, loss of life, loss of wars, loss of democracy. 

In high stakes decision-making we must have confidence that decisions are being made in the interest of the U.K. and our citizens, reinforcing our national security and economic interests. We cannot wholly depend on knowledge, skills, innovation and technology solutions provided by others. We need resilient sovereign capability to deliver decision advantage in a trusted, resilient and sustainable way.