Millions in public funds, but to what end? New study questions UK AI industry focus

A new analysis from the Institute for Public Policy Research (IPPR) reveals that a significant majority of artificial intelligence firms in the United Kingdom are focused on generic business process improvements rather than addressing targeted societal challenges. This finding comes from a comprehensive database covering 3,256 AI companies across the country, marking the first empirical attempt to map the direction of AI deployment in the UK economy.

According to IPPR’s research, only 15 per cent of the examined firms specialise in solving specific problems within defined sectors. Instead, the dominant approach revolves around broad solutions such as analytics, customer service optimisation, and internal efficiency enhancements. While these applications may enhance productivity, the report indicates that their societal impact is limited unless aligned with public needs and government missions.

The UK government has highlighted AI as a driver of public sector innovation, yet the report notes a disconnect between public investment and mission-aligned outcomes. Nearly one in five AI firms have received public funding at some stage, but a large portion of these resources are being channelled into companies with generic or commercially-oriented offerings.

What sectors dominate UK AI deployment, and where are the gaps?

The study finds that 70 per cent of AI activity is concentrated in knowledge-intensive industries, such as professional services, finance, healthcare, and communications. Within this cohort, software engineering dominates, with AI often deployed as a coding assistant. Customer operations, marketing, and supply chain logistics also receive significant attention. In contrast, innovation in areas such as public health prevention, inclusive transport, and environmental monitoring remains relatively scarce.

In healthcare, for example, AI-driven innovations predominantly focus on diagnostics, drug discovery, and treatment improvement. Kheiron Medical Technologies uses machine learning to enhance breast cancer detection, while companies like Isomorphic Labs and Pear Bio are streamlining drug development and treatment personalisation. However, only 12 per cent of healthcare AI initiatives target prevention, a key area for improving population health outcomes and reducing long-term costs for the NHS.

Similarly, in the transport sector, most AI applications relate to autonomous vehicles, logistics, and operational efficiency. IPPR finds that only 9 per cent of firms focus on improving transport accessibility, integrating multimodal transit, or enabling demand-responsive systems. These use cases are crucial for aligning transport innovation with net zero goals and broader economic inclusion.

Are UK AI firms creating new models or relying on existing ones?

The report highlights that most UK AI companies do not develop proprietary models, instead relying heavily on third-party providers such as OpenAI, Meta, or Anthropic. Less than half of firms explicitly state that they create their own foundational AI models. This reflects the growing infrastructure costs and complexity associated with training large-scale models, which have risen substantially since 2020.

This dynamic shifts the emphasis of UK AI innovation from foundational research to application-layer development. Firms are increasingly building services and products on top of established models, focusing on implementation rather than novel algorithmic breakthroughs. For example, PwC UK has integrated OpenAI’s models into its tax advisory services, while Octopus Energy employs generative AI to handle customer service interactions at scale.

How can the government shape the UK AI industry towards public value?

The IPPR proposes a four-pronged strategy to realign AI deployment with national missions and public value creation. The recommendations include establishing a dedicated AI Tracking Unit within the government to monitor the types of AI being deployed and identify existing gaps. This unit would support strategic forecasting of AI’s impact on employment, productivity, and public service delivery.

A second recommendation involves breaking down high-level government missions—such as improving healthy life expectancy or decarbonising transport—into concrete, solvable problems. These targets would enable innovation agencies to design funding calls that are more specific and outcome-oriented. Australia’s CSIRO model is cited as a precedent, where large missions are operationalised through detailed problem statements and public-private collaboration.

Third, IPPR advocates for aligning funding policies—through Innovate UK and the British Business Bank—with clear mission outcomes. Rather than supporting broad sectoral innovation, these agencies could prioritise ventures addressing defined challenges in healthcare, energy, mobility, or education. Embedding mission-oriented metrics into grant assessments and procurement criteria would ensure consistent alignment with government objectives.

Lastly, the report emphasises the use of demand-side incentives such as outcome-based procurement and adoption subsidies. Procurement policies that prioritise performance over rigid specifications could stimulate innovation in under-served areas. The 2025 Procurement Act offers new flexibility, potentially allowing public sector buyers to take a more agile and experimental approach to acquiring AI-enabled solutions.

What role does innovation policy play in fostering meaningful AI adoption?

Historical lessons from general-purpose technologies such as electricity and computing suggest that their full societal impact often emerges from deployment innovations rather than initial breakthroughs. The report notes that AI is likely to follow a similar trajectory, where process innovations—supported by tailored digital infrastructure and workplace adaptation—drive large-scale productivity and welfare gains.

However, this also means that AI’s benefits are not automatic. Without policy intervention, deployment will likely continue to cluster in sectors where financial incentives are strong but public value is marginal. The IPPR warns that AI accelerationism, without strategic direction, risks deepening inequalities and missing opportunities for transformative change.

The institute suggests that new mission boards should be given the authority to coordinate across departments, set procurement goals, and oversee funding streams to ensure consistency in objectives. These boards could enable a feedback loop between AI developers, regulators, and government departments, fostering an ecosystem that incentivises long-term innovation over short-term efficiencies.

Why does this matter for the future of UK innovation and competitiveness?

With the UK already leading Europe in generative AI patent filings and home to significant AI talent, the country stands at a critical juncture. IPPR’s analysis points out that while AI activity is widespread and dynamic, the direction of this innovation is unclear. Without a coherent strategy to steer development toward national goals, the UK risks failing to capitalise on its potential as a global AI leader.

By combining public investment, procurement reform, and strategic coordination, the government could reshape the AI sector into one that not only drives growth but also addresses pressing societal challenges. The report argues that aligning innovation with missions—rather than relying on the market alone—will be essential if AI is to deliver sustainable, inclusive, and equitable outcomes in the years ahead.