Insurance is about covering the risks of the unexpected.
So a company losing money can be insured but the insurance will only cover the costs of unexpected events (say, the business premises burning) rather than the normal business (which in this example actually loses money).
My understanding is that Insurance doesn’t want to cover things done via AI because AI makes lots of mistakes and has a much higher probability of making massive mistakes (the kind of mistakes that can kill people or destroy companies) which even untrained humans will not do - AI will advise a suicidal person to kill themselves or somebody wanting to do pizza to put glue on top, whilst even the most untrained person will not: it’s a bad idea to have an untrained person give psychological advice but it’s an even worse idea to have an AI do it.
In other words AI in the loop increases the probability of exceptionally bad unusual situations happening, so the Insurer doesn’t want to insure companies which use it because that Insurer would have far more and far bigger claims from such a client.This is the same as an insurer refusing to insure companies that use untrained personnel to do really high risk things or which place themselves in high risk situations just to save a buck (for example, storing highly flammable materials in an area with lots of sparks or fire use).
It’s pretty standard for an insurer to refuse to provide insurance to operations which are taking huge unnecessary risks to save money and AI use definitely is a huge unnecessary risk in lots of situations (for example, lawyers using AI risk being disbarred if constantly in submissions to judges the AI references non-existing case law which it invented, same thing it would happen if they themselves constantly did something like that).
Insurance is about covering the risks of the unexpected.
So a company losing money can be insured but the insurance will only cover the costs of unexpected events (say, the business premises burning) rather than the normal business (which in this example actually loses money).
My understanding is that Insurance doesn’t want to cover things done via AI because AI makes lots of mistakes and has a much higher probability of making massive mistakes (the kind of mistakes that can kill people or destroy companies) which even untrained humans will not do - AI will advise a suicidal person to kill themselves or somebody wanting to do pizza to put glue on top, whilst even the most untrained person will not: it’s a bad idea to have an untrained person give psychological advice but it’s an even worse idea to have an AI do it.
In other words AI in the loop increases the probability of exceptionally bad unusual situations happening, so the Insurer doesn’t want to insure companies which use it because that Insurer would have far more and far bigger claims from such a client.This is the same as an insurer refusing to insure companies that use untrained personnel to do really high risk things or which place themselves in high risk situations just to save a buck (for example, storing highly flammable materials in an area with lots of sparks or fire use).
It’s pretty standard for an insurer to refuse to provide insurance to operations which are taking huge unnecessary risks to save money and AI use definitely is a huge unnecessary risk in lots of situations (for example, lawyers using AI risk being disbarred if constantly in submissions to judges the AI references non-existing case law which it invented, same thing it would happen if they themselves constantly did something like that).