Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsAI extinction warnings became much more visible after ChatGPT’s public launch in late 2022 and a burst of high-profile statements and policy debate in 2023. But “everyone” does not think AI is about to kill us: surveys find a mix of concern, optimism and uncertainty, and they ask about very different things—from everyday AI use to human extinction.
Why did AI-doom warnings suddenly become so visible?
ChatGPT made AI a mass public experience
Generative AI moved from a subject many people encountered mainly through specialist or policy discussions into a tool that millions could try directly. ChatGPT’s launch in late 2022 made AI’s capabilities—and its mistakes—visible in ordinary work and conversation. That did not create the debate about advanced AI risk, but it gave the debate a much larger audience.
As an Amazon Associate I earn from qualifying purchases.
Warnings and policy arguments arrived together in 2023
In spring 2023, a Future of Life Institute letter called for a pause on training systems more powerful than GPT-4. Around the same time, the Center for AI Safety published a statement saying: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” The statement was made by the organization, not by a single author. Prominent industry leaders signed or discussed severe risks, bringing the issue further into public and policy debate.
The Atlantic’s 2025 account describes extinction-risk warnings entering mainstream discussion in 2022, public attention waning after the initial shock, and renewed attention in 2025. That is a reported interpretation of the chronology, not a study proving what caused the shifts in attention. The debate also predates ChatGPT; the change was in visibility, not the sudden discovery of the underlying concern.
#1 Best Overall
Do surveys show that people think AI will cause human extinction?
No single poll answers that question for everyone. The studies below differ in who was surveyed, when they were surveyed and what they were asked. Their percentages should not be treated as measurements of the same belief.
| Study and population | What it measured | Finding and how to read it |
|---|---|---|
| Pew Research Center, U.S. adults (5,410) surveyed August 12–18, 2024, and U.S.-based AI experts (1,013) surveyed August 14–October 31, 2024; reported in 2025 | Whether respondents were more concerned than excited about increased AI use in daily life, and whether AI would have a positive impact on the United States over 20 years | Among adults, 51% were more concerned than excited, versus 15% of surveyed AI experts. On the separate 20-year impact question, 17% of adults and 56% of experts expected a positive impact. These are views about AI generally, not estimates of extinction risk. |
| Pew Research Center, U.S. adults, 2025 report | Concerns about particular effects of AI | Fifty-six percent were extremely or very concerned about AI eliminating jobs, and 66% were highly worried about people getting inaccurate information from AI. The report also found concern about loss of human connection and government oversight. |
| UK Department for Science, Innovation and Technology, 2023 survey across nine countries | Independent safety testing and several possible harms | Agreement that powerful AI should receive independent safety testing ranged from 59% in Japan to 76% in the UK and Singapore. In nearly all surveyed countries, more than half were worried about AI-enabled cyberattacks, help designing biological weapons and humans losing control. This does not mean respondents thought extinction was imminent. |
| Sentience Institute, U.S. AIMS supplemental survey, May–July 2023 | Views on AI safety, development speed and human extinction | Some 72.4% agreed AI safety is one of the world’s most important issues, 48.9% thought AI development was moving too fast, and 51.5% expressed concern that AI could cause the end of the human race. The survey’s wording and sample design matter when comparing these results with other polls. |
| Rethink Priorities, U.S. adults, June 2023 | Agreement with and support for the Center for AI Safety statement | Fifty-nine percent agreed with the statement and 58% supported it. The poll used an opt-in online sample; judgments about probabilities can also be sensitive to question wording. |
| Rethink Priorities, U.S. adults, poll fielded April 14, 2023 | Regulation, a research pause, everyday worry and perceived extinction threats | Under that poll’s framing, 70% favored AI regulation akin to the FDA and 51% supported a pause on AI research, while 72% reported little or no everyday worry about AI. Asked which of five listed existential threats was most likely to cause human extinction, 4% selected AI; it ranked below four other threats. Those answers show why support for safeguards is not the same as expecting catastrophe. |
| Forecasting Research Institute, 2023 adversarial collaboration | Forecasts of AI existential catastrophe by 2100 from two deliberately opposing groups | The median forecast among skeptics was 0.10% at the start and 0.12% at the end; among concerned participants it was 25% at the start and 20% at the end. The study shows sharp disagreement among its selected forecasters, not a representative estimate of all AI researchers. |
These figures are not a 2026 polling snapshot. The Pew comparison is based on fieldwork in 2024, reported in 2025; the other polling and forecasting results summarized here are from 2023. They do not establish how public opinion changed during 2025 or 2026.
What do people mean when they say AI could “kill us”?
The phrase bundles together present-day problems and uncertain, high-consequence future scenarios. They differ in mechanism, evidence and timescale; evidence for one should not be used as proof of the other.
Free tools Windows power users keep installed
One-click scans. No signup required.
Harms people can encounter now
People raise concerns about inaccurate AI-generated information, disruption to jobs, impersonation and privacy. Pew’s 2025 findings illustrate that job loss and inaccurate information are prominent concerns among U.S. adults. These harms matter in their own right, but they do not demonstrate that AI will cause human extinction.
Misuse and loss-of-control scenarios
Other worries involve AI helping someone carry out cyberattacks or design biological weapons, or future systems becoming difficult for people to control. The nine-country survey measured worry about some of these possibilities. Extinction-risk arguments focus on the most severe outcomes, but their likelihood depends on future capabilities and circumstances that remain uncertain. A present-day AI error is not proof of such an outcome; uncertainty about advanced AI does not make current harms disappear.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why do informed people disagree so sharply?
They are often answering different questions. One person may be judging whether today’s AI causes enough harm to warrant safeguards; another may be estimating the chance that much more capable systems could eventually produce catastrophe. A survey asking whether someone supports safety testing cannot tell us their probability estimate for extinction. Nor does concern about a possible outcome mean a respondent thinks it is likely or imminent.
Rank #4
The Forecasting Research Institute’s 2023 collaboration makes the disagreement unusually visible: its selected skeptic and concerned groups gave sharply different median probabilities for catastrophe by 2100. Because the participants were deliberately recruited for opposing views, those numbers demonstrate a divide among the study’s groups, not a consensus or a representative expert forecast.
Recommended Free Tools
Sampling and wording matter too. Rethink Priorities cautioned that its June 2023 opt-in poll could be sensitive to how probability questions were phrased. Sentience Institute’s results came from a different survey and sample design. Their percentages can describe what respondents said under those conditions, but they cannot be lined up as if each poll asked the same representative group the same question.
Quick Recap
What is a sensible takeaway?
- The public prominence of AI extinction warnings rose around late 2022 and 2023; the debate itself is older.
- “Everyone thinks AI will kill us” overstates the evidence. People express concern about specific harms, support safety measures, and hold sharply different views about existential risk.
- In Pew’s 2024 fieldwork, surveyed AI experts were less likely than U.S. adults to be more concerned than excited about increased AI use in daily life, and more likely to expect a positive national impact over 20 years. Neither question measured belief in extinction.
- Current harms and future catastrophe scenarios deserve separate treatment. Taking present problems seriously does not settle the likelihood of an extreme future risk, and uncertainty about that risk is not a reason to dismiss present problems.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
