UK prioritises trusted AI access over a homegrown frontier model
Britain does not need its own frontier AI model to compete with the US and China if it can secure access through trusted allies, AI minister Kanishka Narayan has said. Speaking during a Silicon Valley visit, he argued that limited capital, talent and resources should support areas where Britain can establish a competitive advantage, rather than replicate overseas capabilities.
Narayan said the government would support successful domestic frontier-model developers but would not commit substantial public funding to that goal. He pointed to the AI Security Institute’s access to advanced models before public release as evidence that Britain can retain influence without building equivalent technology. Chip design, compound semiconductors and scientific research are priorities for developing capabilities that international partners could depend on.
The approach faces pressure from industry leaders concerned about reliance on American technology companies. Trade body UKAI, representing more than 270 businesses, has called for a comprehensive sovereign AI strategy. Its chief executive, Tim Flagg, argued that commitments including the government’s £500m Sovereign AI Fund lack a coherent plan for building domestic capabilities. Narayan framed the government’s objective as strengthening British leverage, control and agency rather than duplicating allies’ work.