TL;DR
The CBI, with Oliver Wyman, wants the next ten years designated Britain’s “adoption decade” and AI placed at the centre of economic and workforce strategy. Its central finding is a widening execution divide: firms scaling AI across the organisation are pulling away from those still running pilots. Not everyone quoted agrees on the urgency.
The number behind the argument
Work the Oliver Wyman Forum carried out with the NYSE puts it starkly. Among firms leading on AI use, 49% are meeting or beating the return they expected. Among those lagging, the figure is 15%. Same technology, three times the hit rate.
The CBI’s four asks follow from that. It wants a business AI adoption delivery group, a voluntary code and public pledge on responsible adoption, investment in skills at every level underpinned by a national AI literacy standard, and regulation firms can navigate without specialist help.
Chief executive Rain Newton-Smith framed the shift as one of execution rather than invention: the next chapter, she argued, will be defined by how fast Britain can put AI to work at scale, which means “moving beyond pilots” and giving workers “the skills and security to embrace change”.
A dissenting note in the same report
The government’s response was warm. AI minister Kanishka Narayan, appointed by Prime Minister Andy Burnham, called it a welcome contribution and restated the target of AI skills for ten million workers by 2030, backed by more than £200m.
Professor Simon Johnson, who chairs the UK’s AI Economics Institute, struck a different tone on Newsnight. The transformation will be intense, he said, but “the companies are exaggerating” — they are trying to impress investors — and Britain “has some time” to prepare. That is an awkward fit with a report built on the risk of being left behind.
Looking forward
Separate Lloyds research released the same day found 61% of firms already using AI but 31% reporting a skills gap, and adoption tracking company size closely. That is the CBI’s execution divide restated as a size divide. A voluntary code and a literacy standard are cheap instruments; whether they reach a sub-£1m-turnover business that does not yet believe it needs AI is the question neither document answers.