Thank you for sharing this, you’ve communicated many risks clearly and succinctly.
One arena of defense AI that so far has received precious little attention I think is the potential for AI enabled warmaking to substantially degrade humanitarian conventions (such as not attacking civilian water supply, electrical grids, or sanitation facilities).
I hope to contribute some research effort to that issue soon, but I hope others do too, especially in light of the strong likelihood that more advanced long range missile systems making use of computer vision models are already being used by Russian forces against civilian infrastructure in Ukraine.
Thank you for sharing this, you’ve communicated many risks clearly and succinctly.
One arena of defense AI that so far has received precious little attention I think is the potential for AI enabled warmaking to substantially degrade humanitarian conventions (such as not attacking civilian water supply, electrical grids, or sanitation facilities).
I hope to contribute some research effort to that issue soon, but I hope others do too, especially in light of the strong likelihood that more advanced long range missile systems making use of computer vision models are already being used by Russian forces against civilian infrastructure in Ukraine.