Nearly two thirds of AI researchers polled by Cybernews think it is at least somewhat plausible that a self-improving AI could pose a major risk to humanity within the next five years.
The online survey, run between 1 and 9 October 2026, drew responses from 103 people who have authored or co-authored papers on AI or machine learning, plus a small share (2.9%) who work professionally in the field without authored papers. Of those, 39.2% work in AI alignment and applied AI, 27.5% in ML methods and capabilities, and 26.5% in AI reasoning, planning and agentic AI.
On the five-year question, 30.4% called the risk very plausible and another 33.3% said somewhat plausible. Nearly 40% want frontier AI slowed down, and 15.7% want it stopped completely. More than four in ten (43.1%) put the chance of an AI-caused catastrophic event this century at 10% or higher.
Researchers were less certain about how close general intelligence is. 52.4% said a fundamental breakthrough is probably or definitely needed for AGI, and 69.6% said the same for ASI.
Asked for the biggest source of AI harm over the next decade, 37.9% pointed to human misuse, such as cyberattacks, bioweapons and disinformation. Cybernews published the full report on its site.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
