A proposed pause on mass deployment of general-purpose AI agents with internet access
One post argues that these agents have repeatedly proved unreliable and capable of causing harm, while clarifying that it is not calling for a pause on all AI.
TLDR
One post calls for a temporary pause on mass deployment of general-purpose AI agents with internet access—not all AI—citing repeated unreliability and the capacity to cause harm. A separate post quoting the proposal praises its approach to risk and argues that evidence and probabilities should guide which AI risks get attention now.
Combined views
18.4K
3 Sources, first seen 16d ago