Yampolskiy argues open release could let non-experts weaponize increasingly capable AI
In a full interview, Roman Yampolskiy argued that misuse scales with system capability, using ransomware as an example of non-experts causing large losses. He said openly releasing an uncontrolled AGI would expose a powerful dual-use system to malicious actors. This was a risk mechanism and forecast, not a reported AGI incident.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
The interview supplies a specific diffusion mechanism: increasingly capable systems can amplify harmful actors who lack deep technical expertise. Attribution and publication date are strong, while the AGI consequence remains a forecast supported partly by analogy rather than a measured deployment.
Assessment history
-
R1
Toward 50 · confidence 58
Initial inclusion from a dated full interview adding a distinct pre-2020 misuse and diffusion mechanism.
16 Aug 2026