AIThe case for an AI safety layer built into the operating system
A user proposes AI warnings against suspicious links and texts, suggesting the models could probably run on-device.
TLDR
A user argues for an AI security layer built into the operating system that would warn people not to click certain links or answer certain texts. They suggest scam-detection datasets would benefit from network effects and that the models could probably run locally. Revisiting the idea, the same author frames it as “AI protecting us from AI” and considers what “judgment as a service” layers might offer.
Combined views
3.7K
1 Source, first seen 14h ago
The case for an AI safety layer built into the operating system
A user proposes AI warnings against suspicious links and texts, suggesting the models could probably run on-device.
TLDR
A user argues for an AI security layer built into the operating system that would warn people not to click certain links or answer certain texts. They suggest scam-detection datasets would benefit from network effects and that the models could probably run locally. Revisiting the idea, the same author frames it as “AI protecting us from AI” and considers what “judgment as a service” layers might offer.