UK government AI strategy not tied to US models that currently dominate (www.computerweekly.com)

The UK government is taking steps to reduce its dependency on US artificial intelligence models, following revelations that 10 out of 11 AI tools used in central government rely heavily on American suppliers like Google, OpenAI, Microsoft, and Anthropic. This reliance raises concerns over potential control and accessibility issues, prompting the government to adopt a model-agnostic architecture. The aim is to develop AI tools that are flexible and not tied to any single vendor’s ecosystem, allowing for future adaptability and resilience in the face of changing circumstances. UK AI Minister Kanishka Narayan emphasized the significance of maintaining technological sovereignty, asserting that control over crucial technology is vital for national security and democratic power. By implementing a modular API and orchestration layers, the government can safely prototype and evaluate various AI models, fostering a competitive environment for investment without compromising on resilience. The strategy is positioned as a proactive measure aimed at safeguarding the UK's technological future while promoting innovation that is in the best interests of its citizens.