Why Trust Is the Missing Piece in the AI Puzzle

The Promise vs. the Paradox

Artificial intelligence is now firmly on center stage in political speeches and economic roadmaps. Promises of efficiency, progress, and even job creation dominate the discourse. And yet, for many people, AI still feels distant—or worse, dangerous. A new report from the Tony Blair Institute for Global Change, conducted in partnership with Ipsos, makes this tension unmistakably clear: while usage of generative AI is on the rise, public trust remains stubbornly low, threatening to undercut much of the enthusiasm surrounding AI’s future.

What the Data Says

The report offers a striking portrait of the current state of AI adoption in the United Kingdom. More than half of UK adults have used generative AI tools in some form, whether for work or personal use. But nearly half of the population has never interacted with these tools at all. This split in exposure has a profound effect on perceptions of risk. Among those who have never used AI, a full 56 percent consider it a risk to society. In contrast, among weekly users, that concern drops dramatically to just 26 percent. Clearly, familiarity breeds not only confidence but also a nuanced understanding of the technology’s limitations and potential.

Public acceptance of AI is also highly context-dependent. People are far more comfortable with AI being used to optimize traffic or assist in medical diagnoses. However, enthusiasm plunges when AI is used to monitor employee behavior or target political messaging. The use-case, it turns out, is just as important as the technology itself in shaping public perception.

The report also highlights demographic disparities in trust. Younger people tend to be more optimistic about AI, while older generations remain more cautious. Interestingly, those working in sectors likely to be most affected by AI—such as healthcare and education—are among the least confident in its integration. That hesitancy signals a deeper issue: those who understand the stakes best may also be the most concerned.

Why Trust Matters

Public trust is not just a matter of comfort; it is a foundational requirement for the responsible rollout of AI systems. Without trust, public support for AI in critical sectors could falter. This mistrust creates pressure on regulators, who may respond with heavy-handed oversight or even bans, stifling innovation before it has a chance to demonstrate its benefits. Moreover, trust is essential for public adoption. The most promising applications of AI—whether in healthcare, public transport, or education—depend on people actively using and engaging with these systems. If people opt out due to fear or uncertainty, the value of these technologies diminishes.

The trust deficit also threatens to deepen social divides. If certain groups—especially those who are less tech-savvy or more economically vulnerable—feel alienated or misused by AI systems, the result could be a two-tier society where benefits accrue to the few and risks are borne by the many. That dynamic risks not only social unrest but also undermines the political legitimacy of institutions promoting AI.

Toward Justified Trust

The Tony Blair Institute report outlines several key actions to build what it calls “justified trust.” First, the narrative around AI must change. Rather than abstract promises of national productivity or economic transformation, governments and companies should focus on real-world improvements. If AI is seen as a tool that shortens hospital queues or reduces traffic congestion, people are more likely to accept and even welcome its integration.

Second, trust cannot be built on hype alone. Real-world results must be shared transparently. Pilot projects should be evaluated not just on technical performance but on human outcomes. What impact does an AI system have on a teacher’s workload or a patient’s care experience? These are the metrics that matter to the public.

Third, regulation must be visible, specific, and credible. People need to know there are rules—and that those rules are enforced. Governance frameworks should include oversight mechanisms that are sector-specific and responsive to the evolving capabilities of AI systems. Accountability is key. If a system makes a bad decision, there must be a clear line of responsibility.

Fourth, education and upskilling are critical. Many fears about AI stem from ignorance about what these systems can and cannot do. Public education campaigns can demystify AI and help people use it more safely and effectively. These efforts should also target groups that may be at risk of exclusion—such as the elderly or those in low-income communities—ensuring that no one is left behind in the AI revolution.

Finally, public engagement must be meaningful. Citizens should have a voice in how AI is developed, where it is deployed, and under what conditions. This could take the form of public consultations, citizen juries, or transparent policy-making processes. When people feel heard, they are more likely to feel ownership—and more likely to trust.

A Global Challenge

The UK is far from alone in facing this trust deficit. A recent international study on AI attitudes found that while many people use AI tools intentionally, fewer than half trust them. Similar trends are visible in the United States, where citizens express concern about regulatory gaps, ethical oversight, and misuse of personal data. Academic research supports these findings, arguing that accuracy and efficiency are not enough; ethical, legal, and fairness considerations must be at the forefront of AI deployment.

Tensions and Trade-offs

Of course, building trust is not a simple task. There are trade-offs. Transparency and explainability can slow down development. Stronger regulation can impose constraints that reduce agility. And the pace of technological change often outstrips that of legislative bodies. Furthermore, people don’t just mistrust AI—they often mistrust the institutions behind it. Governments, corporations, and tech companies all face crises of credibility. Rebuilding that trust will take time, consistency, and a demonstrated commitment to public interest.

There’s also the matter of global variation. What works in one country may not be applicable in another. Cultural attitudes toward technology, authority, and risk differ significantly. A strategy that builds trust in the UK might not have the same effect in, say, Japan or Brazil.

Looking Ahead

Despite the challenges, there are reasons for cautious optimism. More governments are moving toward comprehensive AI regulation. Some companies are adopting responsible AI principles, conducting bias audits, and opening their models to scrutiny. Early pilot programs in areas like healthcare and public services are starting to yield tangible benefits that may shift public sentiment. But for these efforts to succeed, they must be visible, inclusive, and credible.

The future of AI hinges not just on the brilliance of its engineers or the ambition of its proponents. It hinges on whether ordinary people feel they can trust it. Trust is not a bonus feature—it is the core infrastructure of adoption. Without it, even the most powerful systems will struggle to take root. But with it, AI becomes more than a tool. It becomes a partner in building a better, more equitable society.