Government urged to rethink what ‘sovereign AI’ means

UK technology sovereignty should be defined by the control government retains over critical systems rather than where the technology comes from, according to Red Hat.

The intervention comes amid growing scrutiny of the UK’s reliance on overseas tech providers, particularly as AI becomes more embedded in public services.

The Commons Science, Innovation and Technology Committee warned last month that the government needs a “realistic plan to achieve sovereign capabilities in critical areas” or risk having access to technology cut off “at the whim of its partners”.

The committee said recent US restrictions on some AI models demonstrated the potential risks of relying on allies for access to technologies critical to economic growth and national security. It called on government to set out realistic ambitions for sovereign capabilities in areas including AI, quantum and space.

The FT this week reported that the government is assessing the potential economic and cybersecurity consequences of the UK losing access to frontier AI models.

Dame Chi Onwurah, chair of the Science, Innovation and Technology Committee, told the FT that the government needed a strategy setting out “how an acceptable level of dependency is defined”.

But Jonny Williams, chief digital adviser for UK public sector at Red Hat, argues that technological sovereignty does not necessarily require government to own or source all its technology domestically.

“Sovereignty is really about agency, choice and control, and less about geography. Acceptable dependency comes down to this: can a public body run, audit, and fix its own infrastructure without waiting on a supplier’s permission? Who owns the technology?” he told Think Digital Partners.

Defining acceptable dependency

The Science, Innovation and Technology Committee said in July that the UK was already in a global race for sovereign technology capabilities, with AI becoming a “central arena” for both competition and collaboration.

Its report argued that government currently lacks a coherent strategic framework for using the UK’s scientific and technological strengths to support its wider economic and diplomatic objectives.

For Williams, however, reducing dependency does not necessarily mean eliminating overseas suppliers. Instead, public bodies need to understand where dependencies exist and whether they retain meaningful control if circumstances change.

“Organisations must evaluate what they depend on, under what conditions, and for how long,” he said.

Open source and government control

Williams argues that open source can give public bodies greater control over those underlying dependencies.

“Standardising on enterprise open source keeps the code in a public body’s hands even if a vendor changes its terms or restricts access,” he said.

“This logic applies from the AI layer through to the operating systems and cloud platforms underneath it.”

The parliamentary committee has warned that the UK cannot necessarily assume continued access to critical technologies from allies as geopolitical competition increases.

Onwurah said when the committee published its report that there was “a global race for sovereignty in technologies like AI, whether the government recognises it or not”.

She added that relying on the UK’s international leverage might not be enough to maintain access to those technologies.

Williams maintained that open source can provide organisations with transparency into the tech they use while retaining greater freedom to modify it or move between suppliers.

“Open source provides transparency and the ongoing ability to innovate, modify the tech, or switch providers, without being locked into decisions you can’t own,” he said.