Public opinion is turning against AI: Why billions in tech projects are being killed by angry voters

Public opposition to artificial intelligence is growing fast in Western democracies, and policymakers have yet to find effective ways to respond. The Economist reports that protests against data centres have already blocked nearly $100 billion worth of projects in the United States, while around 40% of American voters say they want AI banned from most industries.

The backlash is showing up in politics and labour disputes alike. AI megadonors have poured tens of millions into a single congressional race in Manhattan. Workers at Samsung in South Korea have threatened strikes over AI-related pay demands. Yet AI still ranks only 29th out of 39 issues in American voter surveys, suggesting the political fights ahead will be far more intense than those seen so far.

What is driving the anger

Opposition to data centers illustrates the broader mood. More Americans say they would accept a nuclear reactor nearby than a data center. The Economist argues this hostility is partly a consequence of AI leaders themselves repeatedly warning of mass job losses and existential risks from the technology.

The article warns that the backlash itself carries serious risks. AI could drive productivity growth, accelerate medical research, and improve education and green technology. Blocking it through regulation or public pressure could cause lasting economic damage, particularly if authoritarian states such as China push ahead while democratic nations hold back. The article draws a parallel with mRNA vaccine research, which stalled after a pandemic-era public backlash.

Four steps to manage the transition

• Spread benefits widely: Data center companies should fund local communities, and governments should create mechanisms giving citizens a direct economic stake in AI progress, including wage insurance for displaced workers.
• Regulate where it matters: Risks such as AI-enabled cyberattacks and bioterrorism need serious, ideally international, regulatory attention. Targeted rules can undercut arguments for blanket bans.
• Measure everything: Better statistics on AI’s actual effects on employment and energy use would help counter misinformation. The article points to Britain’s AI Security Institute and new AI Economics Institute as potential models.
• Use AI in public services: Governments that deploy AI in healthcare, education, and tax administration can build public trust by demonstrating tangible benefits.

The core argument is simple: making AI models better is no longer enough. Convincing people that the technology works in their interest has become equally important.