AI Investment Must Build Capability, Not Dependency

Artificial intelligence now appears in budgets as infrastructure, software, training, consulting, data, cloud capacity and organisational transformation. Leaders may feel that delay itself is the greatest risk. Yet an institution can spend heavily on AI and become less capable: more dependent on one provider, less able to explain its decisions, more exposed to recurring costs and less invested in the people who understand its mission.

Buying a licence is adoption. Building the ability to choose, use, evaluate, correct and leave a system is capability. The two may travel together, but they are not identical.

A responsible investment pathway is: public or institutional purpose → measured baseline → necessary capability → least-dependent design → people and infrastructure → bounded technology → verified outcome → distribution and full cost → resilience → revision or exit.

The central question is not how much an institution has spent on AI. It is what durable human and institutional capacity the spending leaves behind.

The International Monetary Fund's October outlook, released on 13 October, forecasts global growth of 3.1 per cent in 2026. Reuters reports that strong investment in AI infrastructure is a major reason for the slight upgrade, benefiting advanced economies building platforms and countries supplying hardware. The same outlook finds that only thirty-two countries now account for 95 per cent of global growth, compared with fifty-four countries in April.…