Public scepticism is not a side issue for national AI plans; it can determine whether people use the services and tools governments expect to drive growth. Research from the Tony Blair Institute for Global Change and Ipsos identified lack of trust as the leading reason people avoid generative AI.
The findings also resist a simple story of universal fear. More than half of respondents had tried generative AI during the previous year, and frequent users were markedly less likely to see it as a societal risk. Attitudes changed again when people considered specific applications rather than AI as an abstract category.
The report was published amid strong political interest in AI-led growth, making its emphasis on trust especially relevant. Adoption targets can create pressure to minimise public concern, but acknowledging different reactions to beneficial, intrusive, and high-stakes uses provides a more credible foundation for policy.
Familiarity Changes Perceptions of Risk
Among people who had never used AI, 56 percent viewed it as a risk to society. That fell to 26 percent among weekly users, suggesting that direct experience can replace imagined capabilities with a more grounded understanding of both usefulness and limitations. That distinction matters because the headline capability depends on the less visible work of moving, checking, and interpreting information before a useful output reaches anyone. Tracking that chain shows whether latency has genuinely fallen or has merely been transferred to preparation and review.
Exposure alone is not a complete policy because a bad first experience may deepen distrust. Public training and service trials should therefore be low-risk, voluntary where possible, and designed to show where the system helps as well as where human judgement remains necessary. Security and governance work best when designed into the workflow, with access limits, logs, and clear escalation paths. Retrofitting them after adoption usually leaves blind spots and unclear accountability. Routine access reviews and incident exercises can show whether written controls remain usable under time pressure.
Age and Profession Shape Readiness
Younger respondents were generally more optimistic, while older people expressed greater caution. Technology professionals felt better prepared than workers in healthcare and education, even though those public-facing sectors may experience significant changes in daily practice. The figures describe an announced or observed position at publication time, so deployment quality and repeatable results remain more informative than the scale of the claim alone. Later evaluation should preserve the original assumptions instead of silently judging the proposal against information unavailable when it was made.
A national skills programme should not assume one standard learner. Training needs to reflect job context, confidence, accessibility, and the consequences of error, while giving workers a role in deciding how tools alter responsibilities and service standards. A phased approach creates evidence while keeping reversal affordable. It also gives teams time to strengthen data quality and operating skills before additional autonomy increases the consequences of a mistake. Expansion gates should be tied to performance and readiness, not simply to a date promised in an initial roadmap.
Acceptance Depends on the Purpose
Respondents were more comfortable with AI used to reduce traffic congestion or accelerate cancer detection than with workplace surveillance or targeted political advertising. The contrast indicates that people judge the objective, power relationship, and personal consequence, not merely the underlying technology. Seen operationally, the change touches architecture, staffing, procurement, and support at the same time; it cannot be evaluated as a model upgrade in isolation. Each affected team needs a defined role before launch, particularly when a fault crosses organisational or supplier boundaries.
Use-case distinctions should shape regulation and procurement. A model that recommends traffic timing is not socially equivalent to a system evaluating an employee, so oversight, consent, explanation, and appeal requirements should rise with the effect on rights and opportunities. A credible rollout would define a baseline, track exceptions, and name the person empowered to pause or narrow the system. That makes improvement visible without treating early momentum as proof that every use case should expand. Review dates should be agreed before enthusiasm or sunk cost makes a later decision unnecessarily difficult.
Justified Trust Requires Visible Evidence
The report urged governments to explain benefits through concrete experiences such as shorter hospital waits, easier public services, or less congestion instead of relying on abstract GDP forecasts. It also called for capable regulators and broad access to training. The source therefore points to a system-level change in which interfaces and operating practices are just as consequential as raw technical performance. Compatibility, documentation, and recovery procedures determine whether that performance remains available during ordinary work.
Public-sector deployments should publish service measures and significant limitations in language people can use. Demonstrating that a system improved outcomes, not just administrative efficiency, gives citizens a reason to believe adoption serves them rather than being imposed on them. Teams also need version records and representative tests, because data, models, user behaviour, and external rules will not remain fixed. A result that cannot be reproduced is a weak basis for an enduring process. Change logs should connect each release to observed gains, regressions, incidents, and any revised operating limits.
Procurement terms should support audit, data export, incident assistance, and orderly transition so accountability survives changes in technology or supplier. Trust cannot be demanded through optimistic messaging. It is earned when people can see a useful purpose, inspect evidence, challenge a harmful outcome, and believe that institutions will enforce meaningful limits.
