Hendrycks argues competition may select for less controllable AI
Hendrycks argues that commercial and national competition can reward increasingly autonomous systems, weaken oversight, select against systems that are easy to stop, and create dependence on AI embedded in critical functions.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
The article adds a distinct mechanism for loss of human control: economic and strategic selection can favor autonomous, hard-to-stop systems even without malicious intent, though the long-run selection effect remains an argued forecast rather than a measured outcome.
Assessment history
-
R1
Toward 20 · confidence 55
Initial inclusion from a fully dated first-person article adding a distinct competitive-selection mechanism.
13 Aug 2026
Share this page
-
DoomBench assesses “Hendrycks argues competition may select for less controllable AI” as evidence moving toward doom, with magnitude 20 and confidence 55 out of 100 in the competitive race category.
-
The DoomBench assessment of “Hendrycks argues competition may select for less controllable AI” is based on reporting from TIME and records the editorial rationale, source quality, attribution, and revision history.
-
DoomBench summarizes “Hendrycks argues competition may select for less controllable AI” as follows: Hendrycks argues that commercial and national competition can reward increasingly autonomous systems, weaken oversight, select against...
https://www.doombench.com/news/hendrycks-argues-competition-may-select-for-less-controllable-ai-2023-05-31