New research reveals rising levels of AI anxiety across the US

The role of artificial intelligence is expanding rapidly, significantly impacting the human psyche. According to a new report, people are experiencing serious anxiety regarding AI’s future influence on their lives.

A recent study from Stanford University’s 2026 AI Index reveals that more than half of respondents feel uneasy about AI products, despite the excitement surrounding the technology observed in recent years. The report suggests that public concern stems from AI’s potential to negatively impact everything from jobs and economy to elections and personal relationships.

The report further states that AI safety is failing to keep pace with rapid developments, noting that incidents have more than tripled since ChatGPT launched in 2022. Recent research underscores a technical conflict were improving one user aspect of AI like safety can inadvertently weaken another, such as accuracy. This has led to a rising trend of skepticism; Gallup data shows Gen Z’s excitement plummeted from 36% to 22% in just one year.

I think a lot of AI leaders are just out of touch with normal people and don’t realize that fears of skynet are not what is primarily driving anti-AI sentiment,” told US-based behavioral scientist Caroline Orr Bueno. “That exists, obviously, but most people are way more concerned with their paycheck and the cost of utilities.”

Escalating tensions surrounding AI are sparking direct actions against the companies developing it. Online groups have gained traction in popularity in recent months, with some taking extreme measures to push an anti-AI agenda.

Over the past few days, the residence of OpenAI CEO Sam Altman has been targeted, and subsequent threats have been made against the company’s headquarters. The situation began when San Francisco police reported that an unidentified male threw a Molotov cocktail at a North Beach residence home identified as belonging to Altman.

The report further states that a significant plurality of Americans (41%) feel current federal regulations are insufficient; globally the EU is more trusted to manage AI than either the US or China. AI has achieved market penetration faster than the PC or the internet, but incidents or harms spiked from 233 in 2024 to 362 in 2025. Furthermore, despite AI’s rapid advance, frontier models read analog clocks accurately only around 50% of the time, compared to 90% for humans. Ultimately, the study concluded that responsible AI is failing to keep pace with technical capabilities, as safety benchmarks last even as adoption becomes nearly universal across organizations.