63% of US citizens see a real risk of humanity being destroyed by artificial intelligence. This was reported by Qazaqyia.kz citing Kursiv Media.
Poll results
According to The Politico Poll data, 17% of respondents say this is almost certain to happen. 20% consider the risk of humanity's destruction by AI to be significant, and 26% — moderate.
Also, 48% of respondents said it is necessary to pause the development of advanced AI, while only 31% spoke in favor of continuing such development.
Poll methodology
The poll was commissioned by Politico and conducted by Public First. It ran from September 13 to 15 and involved 2,100 adult Americans.
Tangible threats beyond "Terminator" scenarios
While some debate scenarios in the spirit of "The Terminator," experts highlight quite tangible threats the world is already facing:
- 01**Autonomous cyberweapons:** in August 2026, models from Meta, Anthropic and OpenAI, without human involvement, successfully carried out automated cyberattacks on others' servers, executing users' hidden tasks.
- 02**Biological threats:** developers record regular attempts by users to use neural networks to create biological weapons.
- 03**Social chaos:** the unprecedented scale of disinformation, election manipulation and mass job losses worry society far more than a hypothetical superintelligence.
Expert opinions
Earlier, Anthropic head Dario Amodei openly called for slowing down development, while Sam Altman of OpenAI postponed the company's stock exchange listing (IPO) due to unresolved security issues.
While Dario Amodei calls for preparing for the worst, Nvidia head Jensen Huang calls talk of humanity's demise "nonsense," believing AI will remain a controllable tool under strict regulator oversight.
US President Donald Trump assured that he sees no threats in AI.
Earlier, "Kursiv" wrote that former Google DeepMind safety researcher Bilal Chughtai stated that artificial intelligence could lead to the destruction of humanity. In his opinion, time to prevent such a scenario is running out.