Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
AI extinction warnings became much more visible after ChatGPT brought generative AI to a mass audience in late 2022, then spread through high-profile letters, public statements and policy debate in 2023. That does not mean the underlying debate began then—or that everyone believes catastrophe is likely. The evidence shows a mix of concern, optimism and uncertainty, and it covers risks ranging from inaccurate information today to disputed scenarios involving future systems.
Why did AI catastrophe warnings suddenly become so visible?
The short answer is timing and reach. ChatGPT gave millions of people a direct experience of generative AI beginning in late 2022. In spring 2023, the Future of Life Institute published a letter calling for a pause on training systems more powerful than GPT-4, while the Center for AI Safety issued a statement arguing that reducing AI extinction risk should be a global priority. Prominent industry figures signed or discussed severe-risk warnings, bringing the subject into mainstream news and policy conversations.
The Atlantic’s 2025 account describes extinction-risk warnings entering mainstream discussion in 2022, fading somewhat after the initial shock, and receiving renewed attention in 2025. That is a reported interpretation of the debate’s trajectory, not a study proving that one event caused public concern to rise. More broadly, the public launch of a widely used technology, alarming warnings, and policy discussion reinforced one another.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallThe debate itself predates ChatGPT. What changed was the visibility of AI, the speed of its adoption and the prominence of arguments about what more capable systems might mean. A sudden increase in attention is not evidence that the danger itself appeared suddenly.
#1 Best Overall
Does everyone really think AI might kill us?
No. “Everyone” is a provocative headline, not a description of public opinion. Polls find meaningful concern, but the results depend on what people were asked, who was surveyed and when. Concern about AI in general is not the same as believing it will cause human extinction.
Pew Research Center surveyed 5,410 U.S. adults from August 12–18, 2024, and 1,013 U.S.-based AI experts from August 14–October 31, 2024. In the public survey, 51% said they were more concerned than excited about increased AI use in daily life; among the surveyed experts, 15% said the same. On a different question, 56% of those experts expected AI to have a positive impact on the United States over the next 20 years, compared with 17% of U.S. adults. These are the views of Pew’s defined samples, not every member of either group—and neither question asked respondents to estimate the chance of extinction.
Other surveys asked more directly about catastrophic risk, but they are not interchangeable with Pew’s general-attitude questions:
| Source and population | Question or measure | Reported result |
|---|---|---|
| UK Department for Science, Innovation and Technology, nine countries, 2023 | Agreement that powerful AI should receive independent safety testing | 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. |
| Sentience Institute, U.S. AIMS supplemental survey, May–July 2023 | Views on AI safety and existential risk | 72.4% agreed AI safety is one of the world’s most important issues; 48.9% thought AI development was moving too fast; 51.5% were concerned AI could cause the end of the human race. |
| Rethink Priorities, U.S. poll fielded April 14, 2023 | Regulation, a research pause, everyday worry and the most likely cause of human extinction | 70% favored AI regulation akin to the FDA; 51% supported a pause under the poll’s framing; 72% reported little or no everyday worry about AI. Among the listed existential threats, 4% selected AI as the most likely cause of human extinction. |
| Rethink Priorities, U.S. survey, June 2023 | Agreement with and support for the Center for AI Safety statement on extinction-risk mitigation | 59% agreed with the statement and 58% supported it. The survey used an opt-in online sample, and probability judgments can be sensitive to question wording. |
The figures describe different questions and samples, collected in 2023 or 2024. They are not a single measure of how many people think catastrophe is imminent, nor a current 2026 polling snapshot. In particular, backing safety testing or agreeing that a risk deserves attention does not mean someone thinks extinction is likely.
What does “AI risk” mean in practice?
People often use “AI risk” to refer to problems with very different mechanisms and time horizons. Keeping them separate helps prevent both exaggeration and dismissal.
Present-day problems
Current concerns include inaccurate information generated by AI, disruption to jobs, impersonation and privacy. Pew’s 2025 report found that 66% of U.S. adults were extremely or very concerned about people getting inaccurate information from AI, and 56% were similarly concerned about AI eliminating jobs. The report also found concern about loss of human connection and government oversight.
These are concerns about effects people can encounter as AI is used today. They matter on their own; they do not, by themselves, establish that human extinction is a likely outcome.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Misuse of more capable systems
A separate concern is that people could use powerful AI systems to help carry out harmful acts. The 2023 multinational survey asked about AI-enabled cyberattacks and assistance designing biological weapons, among other issues. Those are misuse concerns: the central question is what a person might be able to do with a system’s help.
Loss of control
Longer-term warnings ask whether future systems could become difficult for people to control. This is not the same claim as saying current chatbots are about to take over. It depends on what capabilities future systems acquire, how they are deployed and whether safeguards and oversight are effective. The polling and forecasting figures below show that informed people disagree sharply about the likelihood of the most extreme outcome.
How much do experts disagree about extinction risk?
There is no consensus probability established by the polls in this evidence set. One illustration of the disagreement is a 2023 adversarial collaboration by the Forecasting Research Institute. It intentionally recruited two small groups with opposing views, rather than sampling AI researchers or the public representatively. Participants forecast the chance of AI existential catastrophe by 2100.
| Selected forecasting group | Median forecast at project start | Median forecast at project end |
|---|---|---|
| Skeptics | 0.10% | 0.12% |
| Concerned participants | 25% | 20% |
The large gap demonstrates disagreement among the collaboration’s selected participants; it does not tell us what all AI researchers believe. Forecasts about a possible event by 2100 are also different from poll answers about everyday worry, general AI benefits or support for safety measures. Treating those percentages as competing measurements of the same thing would be misleading.
Why can people support safeguards without expecting extinction?
People may favor testing or regulation because they want to reduce a risk, even if they think it is unlikely. A low-probability event can still warrant attention when the possible consequences are severe. Conversely, support for a safety statement or a pause does not reveal how likely a respondent thinks catastrophe is unless the survey asks that directly.
The wording matters, too. Rethink Priorities cautioned that its June 2023 opt-in survey’s results could be sensitive to question phrasing, especially probability judgments. Its April poll found substantial support for regulation and a pause alongside little or no everyday worry among most respondents. Those results are not contradictory: people can want rules for a consequential technology without feeling personally alarmed in daily life.
What should you take away from the warnings?
- Separate attention from belief. More headlines and public statements mean the debate became more visible; they do not show that everyone thinks catastrophe is imminent.
- Ask which risk is meant. Inaccurate output, job disruption, malicious use and loss of control involve different evidence and timelines.
- Check what a poll actually measured. Note the population, field dates and wording before comparing numbers. A question about general AI use is not an extinction-risk question.
- Read forecasts with their limits. A selected forecasting group can reveal disagreement, but its medians are not representative public or expert opinion.
- Hold uncertainty in both directions. Current AI harms do not prove an extinction scenario; uncertainty about future systems does not make present harms irrelevant.
The most comparable public-versus-expert figures here come from Pew’s 2024 fieldwork, reported in April 2025. The other surveys summarized above were conducted in 2023. These data explain why concern and debate became prominent, but they do not establish how views changed during 2025–2026.
Quick Recap
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.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →
