It seems this type of AI might cheat and set ε=0 using the following method:
create an “insurance company” subagent (see below for what it is and why creating it has a utility that is >ε).
give the insurance company a few utilons (paper clips) to hold as hostages.
set a rule that all future subagents will be able to communicate the insurance company.
the insurance company will “buy” all the reward that has a probability < ε from subagents at roughly its fair price (its utility)
all subagents won’t ignore events that happen with probability < ε because their very existence is has a probability of 1 to increase their utility (via the insurance company).
alternatively for any ε you set, it will be profitable for the AI to create a new version of itself with the same utility function and ε′=ε/2 than give the new AI all the resources it has and commit suicide.
alternatively for any ε you set, it will be profitable for the AI to create a new version of itself with the same utility function and ε′=ε/2 than give the new AI all the resources it has and commit suicide.
This doesn’t seem true. A ε/2 AI will take risks, looking for higher utility, that the ε AI wouldn’t.
It seems this type of AI might cheat and set ε=0 using the following method:
create an “insurance company” subagent (see below for what it is and why creating it has a utility that is >ε).
give the insurance company a few utilons (paper clips) to hold as hostages.
set a rule that all future subagents will be able to communicate the insurance company.
the insurance company will “buy” all the reward that has a probability < ε from subagents at roughly its fair price (its utility)
all subagents won’t ignore events that happen with probability < ε because their very existence is has a probability of 1 to increase their utility (via the insurance company).
alternatively for any ε you set, it will be profitable for the AI to create a new version of itself with the same utility function and ε′=ε/2 than give the new AI all the resources it has and commit suicide.
This doesn’t seem true. A ε/2 AI will take risks, looking for higher utility, that the ε AI wouldn’t.