“The core challenge of insuring AI risks is a lack of historical data. For insurance or bond markets to work, they will likely need data that is very hard to come by in the AI world. Insurers and institutional investors estimate risk based on data from the past: namely, observed damages from events similar to those they are trying to insure or invest in. This works well when what might happen in the future looks a lot like what has happened before; it also works when past events are sufficiently similar to compose a “data set” from which one can generalize and predict. But such conditions do not always hold. In the early 2000s, for instance, federal policymakers realized that terrorism posed risks that defied these properties. Nonetheless, they sought a way to estimate and at least partially insure against the threats of terrorism. Enter TRIP.”
Gerd Leonhard Futurist Keynote Speaker Author“The core challenge of insuring AI risks is a lack of historical data. For insurance or bond markets to work, they will likely need data that is very hard to come by in the AI world. Insurers and institutional investors estimate risk based on data from the past: namely, observed damages from events similar to those they are trying to insure or invest in. This works well when what might happen in the future looks a lot like what has happened before; it also works when past events are sufficiently similar to compose a “data set” from which one can generalize and predict. But such conditions do not always hold. In the early 2000s, for instance, federal policymakers realized that terrorism posed risks that defied these properties. Nonetheless, they sought a way to estimate and at least partially insure against the threats of terrorism. Enter TRIP.” We Don't Need to Wait for an AI Disaster to Estimate Its Costs https://open.substack.com/pub/aifrontiersmedia/p/we-dont-need-to-wait-for-an-ai-disaster?r=2w97&utm_medium=ios via Instapaper Download file.txtfile.txt
2.19 KB