Artificial intelligence could boost the economy and public sector productivity, but for these benefits to materialise concerns around the safety and reliability of AI tools need to be allayed. This report looks at the rising demand for AI assurance technologies and sets out how the UK can get a head start in this market.
KEY POINTS
- Artificial Intelligence (AI) could bring economic benefits worth trillions of pounds and transform UK public services.
- For those benefits to be realised, users need assurance that AI tools are safe, secure, and reliable. As a result, there is a burgeoning demand for new AI assurance technologies (AIATs) – solutions for AI alignment, security, auditing, authentication, and risk management.
- New modelling suggests that the global AIAT market will reach $276 billion by 2030, but most AIATs currently need R&D investment to operate at scale.
- The UK has a head-start in the AIAT landscape, with a world-leading AI Safety Institute (AISI) and a fledgling AIAT start-ups scene. But uncertainties around regulation, technical priorities and demand mean that the global AIAT market is still very much up for grabs.
- Handled correctly, the UK could capture a massive share of the emerging AIAT market, complementing growth in the UK AI market more generally and supporting national security and regional industrial strategy.
RECOMMENDATIONS
- Establish an inter-departmental group (IDG) to guide AIAT policy.
- Come up with a three-year roadmap for introducing mandatory AI assurance standards, including an MOU with the EU and US for harmonisation.
- Direct £10m of AISI’s existing budget into seed funding for priority AIATs, and grant companies developing said AIATs free access to public compute.
- Commission a study analysing the technical and market barriers to developing and commercialising priority AIATs.
- Invest £50m in “pull mechanisms” (pay-outs contingent on achieving specific technological goals, such as prizes and milestone payments).
This briefing was has been published simultaneously by UK Day One Project and the Social Market Foundation. The authors are researchers at the Institute for AI Policy and Strategy.