In a stark reflection of growing unease, a recent poll reveals that a majority of Americans believe artificial intelligence could one day spell the end for humanity. The survey, conducted by Yahoo News and YouGov, paints a picture of widespread pessimism amid rapid AI advancements. As AI integrates deeper into daily life, from healthcare to entertainment, public sentiment appears increasingly wary of its long-term implications.

The poll, released in November 2025, found that 53% of respondents think AI will ‘destroy humanity’ someday, with only 23% disagreeing and 24% unsure. This sentiment cuts across demographics, though younger adults and those with higher education levels express slightly less alarm. The findings underscore a disconnect between Silicon Valley’s optimism and Main Street’s apprehensions, as tech giants race to develop more powerful AI systems.

The Roots of AI Anxiety

Public fears aren’t unfounded. Experts have long warned about existential risks from AI, including loss of control over superintelligent systems. A report from the Center for AI Safety, as detailed on their website safe.ai, highlights scenarios like bioterrorism or military AI mishaps that could lead to catastrophic outcomes. These concerns echo in the poll, where respondents cited job displacement, privacy erosion, and even apocalyptic scenarios as top worries.

Historical parallels amplify these fears. As noted in a 2023 article from PMC pmc.ncbi.nlm.nih.gov, AI threatens human health and existence through social, political, and economic channels. The poll’s timing coincides with recent headlines, such as a DeepSeek researcher’s warning at a Chinese internet conference, reported by Reuters reuters.com, expressing pessimism about AI’s societal impact despite short-term benefits.

Shifting Public Perceptions

Social media platforms like X (formerly Twitter) buzz with similar sentiments. Posts from users and polling accounts in 2025 highlight a divide: one Gallup poll shared on X in June showed Americans split 49-49 on whether AI is just another tech advancement or a unique threat. Another X post from AI Safety Memes cited stats where 5-to-1 ratios favor slowing AI development, reflecting broader calls for regulation.

Broader surveys reinforce this. A Pew Research Center study from September 2025 pewresearch.org found Americans worried about AI harming creativity and relationships, yet open to its use in data-heavy fields like medicine. Women, in particular, express higher concern—over twice that of men, according to a post referencing Rutger Bregman’s analysis on X.

Economic and Societal Ripples

The economic stakes are high. An article from The Economist economist.com predicts that in 2026, AI’s true impact—boom, bust, or backlash—will emerge, with potential GDP boosts threatened by deglobalization. WebProNews webpronews.com echoes this, warning that tariffs could offset AI-driven growth in 2025.

Environmental concerns add another layer. The United Nations Environment Programme unep.org reports that AI data centers generate massive e-waste and consume fossil fuel-derived electricity, exacerbating climate risks. This ties into public fears, as a Built In article builtin.com lists 15 AI dangers, including inequality and psychological harm.

Industry Responses and Regulatory Push

Tech leaders are responding variably. Elon Musk’s xAI and others face scrutiny for military applications and unregulated developments, as covered in NSS Magazine nssmag.com. A Startup News article startupnews.fyi notes that agentic AI’s productivity gains in 2025 could pose privacy risks in 2026.

Calls for oversight are growing. An X post from ControlAI in November 2025 referenced over 100,000 signatures for banning superintelligence development. Similarly, a 2023 X poll from AI Notkilleveryoneism Memes showed 82% favoring AI slowdown and federal regulation, a sentiment persisting into 2025 per recent posts.

Global Perspectives on AI Risks

Internationally, views align with U.S. pessimism. A Times of India report timesofindia.indiatimes.com quotes a DeepSeek researcher warning of long-term threats. Virginia Tech’s engineering magazine eng.vt.edu debates AI’s ‘good, bad, and scary’ sides, questioning if changes are for the better.

Bioethics enters the fray too. A 2020 PMC article pmc.ncbi.nlm.nih.gov explores AI’s transformation of human relations and self-knowledge, predicting shifts akin to an industrial revolution. These global insights feed into American fears, as evidenced by the Yahoo/YouGov poll’s alignment with international surveys.

Navigating the AI Future

Despite the gloom, optimism persists in targeted applications. Pew’s report notes openness to AI in weather forecasting and medicine, where data precision offers clear benefits. However, the poll’s 70% insistence on human oversight, as shared on X by Rutger Bregman, signals a demand for balanced progress.

Experts urge proactive measures. The Center for AI Safety focuses on mitigating catastrophic risks, advocating for safeguards against misuse. As AI evolves, bridging the perception gap between innovators and the public will be crucial, potentially through transparent regulations and ethical frameworks.

Voices from the Ground

Real quotes capture the zeitgeist. One X user posted, ‘53% of Americans now believe that AI will destroy humanity at some point,’ referencing Collective Action for Existential Safety. Another from Pinna Pierre highlighted the Yahoo/YouGov findings, emphasizing pessimism compared to tech hubs.

In interviews, respondents in the Android Central-covered poll androidcentral.com expressed fears of AI surpassing human intelligence, leading to unintended consequences. This mirrors earlier 2023 CNBC data shared on X, where 55% worried about AI risks to humanity.

Pathways to Mitigation

Mitigation strategies are emerging. Proposals include pausing advanced AI development, as supported by 5-to-1 ratios in 2023 X polls that continue to resonate. Governments are stepping in; the U.S. explores AI safety agencies, per public support in surveys.

Ultimately, the poll serves as a wake-up call. As The Economist forecasts, 2026 could clarify AI’s trajectory, but addressing public fears now—through education, regulation, and ethical AI—may avert the doomsday scenarios many anticipate.