Nabeel Qureshi Replies on AI Harm Emergence Timing
Reply notes frontier model harms emerge gradually rather than instantly.
In reply to Jason Crawford, Nabeel S. Qureshi stated that people expect this stuff to happen instantly but that it happens eventually and takes time. The exchange sits inside a discussion on AI timelines that addresses why frontier models have not yet caused significant cyber harm to infrastructure or individuals. The packet records a generated headline summarizing the point as AI experts saying cyber harm from frontier models takes time to emerge. The source summary notes that participants observe expectations around when such effects appear. No other details or outcomes are shown in the supplied lines.
Combined views
162
1 post, first seen 2h ago
Nabeel Qureshi Replies on AI Harm Emergence Timing
Reply notes frontier model harms emerge gradually rather than instantly.
In reply to Jason Crawford, Nabeel S. Qureshi stated that people expect this stuff to happen instantly but that it happens eventually and takes time. The exchange sits inside a discussion on AI timelines that addresses why frontier models have not yet caused significant cyber harm to infrastructure or individuals. The packet records a generated headline summarizing the point as AI experts saying cyber harm from frontier models takes time to emerge. The source summary notes that participants observe expectations around when such effects appear. No other details or outcomes are shown in the supplied lines.
Reactions
1 postNS 
Nabeel S. Qureshi@nabeelqu@jasoncrawford Yes. People expect this stuff to happen instantly. It happens eventually but takes time
- likes: 2
- replies: 0
- bookmarks: 0
- reposts: 0