Written submission from Dr Narmin Nahidi (PRO0029)

AI, Behavioural Finance, and UK Growth Policy

Date: 13/11/2025

Inquiry: Finance the real economy

Committee: Business and Trade Committee

I am Dr Narmin Nahidi, Assistant Professor in Finance at the University of Exeter Business School. My research spans corporate finance, fintech, behavioural finance, and the economic implications of artificial intelligence for financial decision-making. A core theme in my work is how technological change interacts with managerial behaviour, risk perception, and organisational investment choices, factors that directly shape business confidence, productivity, and long-term growth. I am submitting evidence to support the Committee’s inquiry because AI is becoming an integral part of financial and business decision-making, yet firms’ responses are driven as much by behavioural constraints as by economic fundamentals. These behavioural patterns help explain the UK’s uneven technology adoption, persistent underinvestment, and widening regional productivity differences.

My submission draws on empirical evidence, policy analysis, and established behavioural-finance research to examine how AI adoption, expectations, and regulatory signals influence corporate behaviour. Understanding these mechanisms is essential for assessing how pro-growth reforms, regulatory clarity, and coordinated government action can improve the UK’s competitiveness and support more widespread and responsible adoption of AI across sectors. This analysis is situated at the intersection of finance and the real economy. Decisions about AI investment, responses to regulatory uncertainty, and managerial biases directly influence firms’ capital allocation, productivity dynamics, and long-term economic performance. These financial choices determine whether the real economy experiences sustained innovation, improved regional outcomes, and durable competitiveness.

1. Business Confidence

Recent UK evidence indicates that firms remain cautious about large-scale investment, particularly in technologies such as artificial intelligence. Official statistics show that only around 9% of UK businesses reported using AI in 2023, with a further share indicating plans to adopt it later. This low uptake signals ongoing hesitation, which is closely linked to how firms interpret policy stability and long-term economic signals. When firms cannot anticipate regulatory direction, the perceived risk of committing to AI investment increases. Behaviourally, this reflects ambiguity aversion: decision-makers tend to overweight the costs of uncertainty relative to prospective gains. A second factor shaping confidence is the rising importance of government communication. Evidence from business-confidence surveys shows that firms frequently cite tax burden, regulatory uncertainty, and policy inconsistency as material drivers of reduced investment willingness. When policy signals conflict, managers often postpone long-term decisions. AI models used internally for forecasting also require stable macro and policy inputs; instability lowers model reliability and further discourages commitment. This illustrates an interaction between machine-driven and human-driven risk assessment.

AI itself depends heavily on confidence in institutional frameworks. Firms with stronger management quality and higher productivity are more likely to adopt AI, reflecting both capability and a greater willingness to interpret policy information optimistically. Empirical work shows that these firms are more sensitive to regulatory signals, adjusting expectations more quickly when uncertainty rises. This sensitivity means poorly coordinated communication disproportionately affects precisely those firms the UK needs to drive innovation. Taken together, business confidence is shaped by both economic fundamentals and behavioural dynamics. The UK can strengthen confidence by improving regulatory continuity, ensuring consistent communication, and reducing uncertainty in areas such as data governance, skills, and AI safety standards. By improving predictability, the government lowers the psychological and financial barriers that currently limit AI adoption and broader investment.

2. Pro-Growth Reforms and Investment

AI has become a central determinant of long-term growth potential. Estimates from UK policy reviews suggest that sustained and safe AI diffusion could raise productivity growth materially, although these figures depend on optimistic adoption scenarios. Despite this potential, firms frequently delay investment because they face behavioural and organisational barriers. Present bias leads managers to focus on short-term operational pressures rather than strategic digital transformation. Loss aversion causes firms to weigh upfront costs more heavily than probabilistic future gains. These behavioural mechanisms slow investment even when aggregate returns appear favourable. Empirical evidence shows that the main barriers to AI adoption include difficulty identifying appropriate use-cases, cost concerns, and skills gaps. These are not purely financial obstacles; they also reflect cognitive frictions. Identifying use-cases requires forward-looking thinking and the capacity to evaluate unfamiliar technologies. Skills shortages increase perceived execution risk, which behavioural finance shows can lead to overestimation of failure probability. These frictions imply that growth-oriented reforms must target managerial beliefs and organisational readiness, not only financial incentives.

The UK’s Industrial Strategy can help overcome these frictions, but only if reforms provide long-term visibility. For AI, firms need predictable regulation, investment in digital infrastructure, and credible signals that skill development will continue. Behavioural evidence suggests that clear multi-year commitments reduce hesitation by lowering uncertainty premiums. Without this, firms retain a “wait-and-see” stance even when incentives are available. For reforms to translate into productive investment, they must integrate behavioural considerations into design. Tax incentives for AI investment, targeted adoption support for SMEs, and structured guidance on AI governance can reduce perceived risk. Ultimately, reform that fails to account for behavioural constraints will underdeliver since technology adoption depends as much on expectations and beliefs as on capital availability.

3. Costs of Doing Business

Firms continually weigh the cost of compliance, tax, and regulation against the expected benefits of growth-oriented investment. Business-confidence monitoring indicates that concern over tax burden and regulatory obligations has increased substantially in recent years. These concerns raise firms’ discount rates for long-term investments, including AI, making them more risk-averse. Behavioural finance shows that when decision-makers perceive a high cumulative burden, they become more conservative, deferring innovations that require upfront cost. AI has the potential to reduce operational and compliance costs, but firms do not automatically factor these benefits into their decisions. Behavioural inertia, combined with fear of disruption, often leads firms to treat AI as an additional burden rather than a cost-saving tool. This is particularly visible among SMEs, which tend to anchor on initial expenditure and underestimate long-run efficiency gains. Empirical SME studies show that adopters often achieve significantly higher productivity, but firms that have not yet adopted systematically undervalue these gains.

A major element of the cost burden is regulatory complexity. Even where regulations are well-intentioned, unclear guidance increases firms’ perceived risk of non-compliance. Behaviourally, this activates loss aversion: firms disproportionately fear the possibility of future penalties relative to potential gains from innovation. Clear, consistent, and predictable regulation can reduce this behavioural barrier, making AI adoption more attractive over the medium term. To support growth, policy should focus on lowering both actual and perceived costs. This includes streamlining compliance processes, providing clarity on AI-related rules, and offering targeted financial support for digital adoption. Crucially, the design of these measures must recognise behavioural factors, especially the need to counteract risk aversion and status-quo bias, so that firms respond as intended.

4. Productivity Growth

The UK’s lagging productivity performance is well-documented. AI offers substantial potential to address this gap, with several studies estimating that generative AI and automation could materially raise output per worker under plausible adoption scenarios. However, these gains are contingent on broad and effective diffusion. Firm-level data show that technology adopters tend to have significantly higher productivity than non-adopters, indicating genuine upside. Yet adoption remains uneven because many firms face behavioural and organisational frictions. Behavioural finance offers a clear explanation: firms often exhibit over-reaction to perceived short-term disruption, under-investment in complementary skills, and excessive caution around new technologies. Managers may overweight anecdotal examples of AI failures, leading to probability distortions that slow adoption. Additionally, uncertainty about workforce impact can inhibit investment even when economic evidence suggests net gains. These behavioural mechanisms provide a missing link between UK technological capacity and observed productivity outcomes.

Technology-driven productivity gains tend to concentrate in firms and regions already positioned to take advantage of them. Without intervention, AI adoption will reinforce existing disparities. Evidence from SME studies shows high potential gains, but only among firms that adopt. If smaller or regionally disadvantaged firms remain reluctant due to perceived risk, national productivity improvements will be muted. To close the productivity gap, policy must combine technological initiatives with behavioural interventions. Firms require confidence that AI investment aligns with regulatory, skills, and governance frameworks. Structured support, managerial-skills programmes, and transparent case studies can reduce uncertainty and accelerate adoption. Without these measures, productivity growth will fall short of model-based projections.

5. The Role of Regions and Cities

AI activity is beginning to diversify geographically across the UK, but adoption remains heavily concentrated in certain regions. While some areas have seen rapid growth in AI-related firms, many regions face structural and behavioural barriers. Firms outside established clusters may perceive higher execution risk due to weaker skills networks, fewer technology partners, and lower peer support. Behavioural finance suggests that such firms will discount future benefits more heavily, increasing reluctance to adopt AI. Regional economic development strategies often emphasise infrastructure and housing, which are necessary but not sufficient. Without parallel investment in digital skills, managerial capability, and trusted local institutions, AI adoption may remain concentrated. Behavioural barriers such as risk aversion and low confidence can prevent firms from exploiting new infrastructure. Therefore, regional strategies must pair capital investment with targeted initiatives that reduce uncertainty and build local capability.

Local leadership can play a central role by providing credible signals to firms. When regional authorities actively support AI skills programmes, facilitate peer-learning, and promote examples of successful adoption, firms update their expectations and perceive lower risk. These behavioural shifts are essential because firms in high-uncertainty environments often overweight potential downsides. To ensure balanced growth, regional policy must explicitly address both economic and behavioural dimensions. Funding for digital-skills hubs, local adoption support for SMEs, and structured regional AI-readiness programmes can help ensure infrastructure investment translates into real productivity gains across the country.

6. Better Regulation

Regulation strongly influences firms’ perceptions of investment risk. Where rules are ambiguous or fragmented, firms interpret this as increased uncertainty, which behavioural finance shows leads to higher discounting of long-term returns. For AI, clarity around liability, safety, data governance, and auditability is essential to reduce perceived risk. Model-based evidence indicates that uncertainty about regulatory direction is among the most frequently cited barriers to adoption. Even well-designed regulation can carry high behavioural costs if communication is unclear. Many firms fear potential future liabilities more than they value potential productivity gains, a classic asymmetry driven by loss aversion. This fear is amplified among SMEs, which often lack in-house legal or technical expertise. As a result, unclear rules can suppress innovation even when compliance costs are manageable.

The government’s regulatory reform agenda provides an opportunity to reduce this behavioural burden. Streamlining regulatory processes, increasing transparency in rule-making, and introducing safe-harbour provisions for responsible AI experimentation would reduce ambiguity. Predictable timelines and coordinated communication across departments would further reduce the perceived risk associated with digital investment. A more coherent regulatory approach would improve efficiency and reduce duplication. Most importantly, predictable regulation lowers the option value of waiting and encourages earlier adoption. This shift is essential if the UK is to move from cautious experimentation to meaningful deployment of AI across sectors.

7. Risks and International Context

The global environment for AI and technological investment is highly competitive. Major economies have launched large-scale AI strategies with clear long-run commitments. Firms comparing jurisdictions consider not only tax and regulatory burdens but also policy coherence, clarity of long-term plans, and the credibility of national institutions. Behaviourally, jurisdictions perceived as unstable or inconsistent tend to attract less investment because firms overweight the risk of future policy shifts. Long-run projections suggest that successful AI adoption could deliver substantial cumulative gains to the UK economy, though these outcomes depend heavily on strategic positioning. Without clear commitments to skills, infrastructure, and governance, firms may continue to interpret the UK as a higher-uncertainty environment relative to competitors. This perception increases discount rates applied to UK investment decisions.

Public attitudes toward AI also influence international competitiveness. Surveys indicate that a significant proportion of UK adults view AI as posing risks to the economy and society. Low public trust can create an expectation of future regulatory tightening, which firms internalise when evaluating investments. Behavioural-finance research shows that expectations of future restrictions elevate perceived risk today. To remain internationally competitive, the UK must project consistency in its AI strategy. This includes aligning regulatory development with industrial policy, coordinating international engagement, and ensuring that domestic policies reduce rather than amplify uncertainty. Without a coherent approach, global investors may allocate more capital to jurisdictions with clearer long-term roadmaps.

8. Government Strategy and Coordination

Effective AI adoption requires alignment across multiple government departments. Current arrangements create fragmentation, with overlapping initiatives and inconsistent signals. Behaviourally, firms struggle to interpret policy when relevant actions occur across separate domains skills, data, regulation, industrial policy, and innovation. This complexity leads to cognitive overload and strategic hesitation. Policy reviews highlight that the UK has strong research and human-capital foundations but lacks mechanisms to translate these into widespread adoption. The absence of coordinated signals increases uncertainty, particularly for firms considering long-term digital investment. Behavioural factors amplify this: when decision-makers face mixed messages, they tend to defer investment until clarity improves.

A coordinated government strategy should integrate behavioural insights into policy design. This includes setting clear multi-year commitments, supporting peer-learning networks, and ensuring communication emphasises stability and continuity. Clear articulation of responsibilities across departments would reduce noise and improve firms’ expectations. Finally, coordination mechanisms should incorporate measurement of behavioural readiness, managerial capability, skills availability, and organisational confidence. These indicators are essential for understanding where intervention is most needed. Without addressing behavioural constraints, policy will struggle to convert the UK’s scientific and technological strengths into broad-based adoption and sustained economic growth.

9. Policy Implications

This evidence suggests several policy actions that can strengthen business confidence, support productive investment, and improve real-economy outcomes. First, regulatory communication should prioritise stability and predictability. Clear multi-year commitments in AI governance, skills development, and infrastructure reduce uncertainty premiums and encourage firms to bring forward investment. Second, policy design should incorporate behavioural insights. Firms facing ambiguity, skills gaps, or organisational inertia often underinvest even when financial incentives are present. Targeted adoption support, managerial-skills programmes, and practical guidance on AI use-cases can help address these constraints. Third, regional policy should go beyond infrastructure by supporting local capability and confidence through digital-skills hubs and structured peer-learning networks. Finally, cross-government coordination is essential. Aligning industrial strategy, regulatory development, and skills policy ensures consistent signals that reduce cognitive overload for firms. These actions can help convert the UK’s technological strengths into widespread productivity gains and sustained economic competitiveness.