AI Anxiety: Americans Grapple with Tech’s Rapid Rise

As artificial intelligence surges forward in early 2026, Americans are displaying a blend of fascination and fear. The month has been marked by eye-catching advances: Anthropic unveiled Claude Opus 4.6, a model boasting multi-agent capabilities that allow coordinated “teams” of AI to tackle complex tasks like financial analysis or document processing. OpenAI countered with GPT-5.3-Codex and its Frontier platform, designed to embed AI agents seamlessly into enterprise systems. Meanwhile, a $200 million tie-up between Snowflake and OpenAI promises to make “agentic AI”—systems that act autonomously on goals, more accessible for businesses handling vast data troves. These developments, alongside China’s Zhipu GLM-5 topping open-source benchmarks and MiniMax’s cost-efficient models, underscore a shift toward practical, workflow-integrated AI. Yet, beneath the hype, public sentiment reveals unease about this accelerating transformation.

In our latest Geostrategy poll in February, engagement with AI is on the rise but tempered by scepticism. Roughly a quarter of respondents, 24% use it regularly, often for tasks like content generation or basic queries, while 39% have experimented sporadically. Another 20% have only seen it in action without trying it themselves, and 11% report no exposure at all. Trust, however, is limited: just 35% express a fair amount or great deal of confidence in AI, with 58% showing little to no trust. The economic outlook tilts negative, over a third (34%) view its impact as more harmful than beneficial, and only 17% see mostly positive effects. A full 61% anticipate a net decrease in jobs due to AI, while just 9% expect increases. On corporate spending, 54% believe companies are overinvesting, compared to 22% who think it’s the right amount.

This caution isn’t isolated. In our past polling, we’ve consistently found that around six in ten Americans think AI is advancing too quickly, with similar shares lacking confidence in government oversight a sentiment that aligns with broader concerns over rapid deployment. We’ve also seen about three-quarters calling for government action to prevent job losses, and over half predicting AI will exacerbate inequality and polarisation.

February’s events likely amplify these fears: the push for agentic systems, capable of independent actions like managing supply chains or drafting legal documents, raises questions about accountability. For instance, if Claude’s multi-agent teams err in a high-stakes financial model, who bears the blame? Such scenarios resonate with public anxieties over job displacement, especially in sectors like manufacturing or administrative work, where AI’s efficiency gains could automate roles en masse echoing our findings where 69% express concern about impacts on factory jobs.

Demographic nuances add depth to the picture. Men and younger adults show higher adoption and optimism. Our poll indicates men are more likely to trust AI (41% fair/great deal vs. 29% for women) and use it routinely (28% vs. 19%), perhaps drawn to its coding and analytical prowess, as seen in GPT-5.3’s enhancements. Those over the age of 65, conversely, report lower exposure (24% never used/seen) and greater wariness (31% not at all trusting), aligning with past findings where older groups prioritise safety regulations. Ideological splits emerge too: conservatives express slightly more trust (39%), possibly viewing AI as a tool for innovation and efficiency, while liberals fret over inequality amplification a concern backed by our earlier data showing 73% favouring interventions against social divides.

Our polling suggests Americans aren’t rejecting AI outright as usage is climbing to 70% having tried it, but demand safeguards to ensure benefits outweigh risks. As agentic tech matures, bridging this trust gap will be crucial.

Download the data tables for this poll.