Containing AI and a push-your-luck rule.
Posted: Sun Jul 04, 2010 12:18 pm
I'm soliciting thoughts on a good push-your-luck rule for player controlled AIs.
In this setting census would be a limiting factor for colonization/expansion. However, AI can be used by the player as "manufactured census" to drive the system EP output equation.
AI would be available in certain levels. The lower levels translate to units of census at some fractional value of 1. Higher levels might be better than 1.
Available AI levels could correspond to tech levels, or, as soon as AI is possible all possible levels could be available. In the latter case, "dumber" AIs represent AIs with more constraints.
Constraints are what keep the AI from going SkyNet on the player. Higher levels of AI would produce higher risks of singularity even as they provide more productivity. Chance of risk might be weighed each year or even each turn (with higher AIs).
Consequences of singularity would be driven by a table with a range of consequences. Most likely: AI burnout / AI suicide--loss of all census/productivity. Less likely -- AI tumor (AI dumbly works to convert all system resources into some random build order and/or AI census). Worst case -- AI effectively becomes a new CM controlled faction with randomly generated AIX traits; roll on table for disposition towards parent race / life in general / or unique goals that AI faction will work towards.
Singularity might be mitigated with player investment in some version of "reaction control rods" that could shut down an AI headed toward singularity (better tech = better chance) with the associated cost of a turn or more's lost productivity.
I would like to see a balanced rule that challenges players to carefully weigh their greed for resources against the risk of creating Cylons.
Thanks for any/all thoughts on design/implementation.
Cheers,
Josh
In this setting census would be a limiting factor for colonization/expansion. However, AI can be used by the player as "manufactured census" to drive the system EP output equation.
AI would be available in certain levels. The lower levels translate to units of census at some fractional value of 1. Higher levels might be better than 1.
Available AI levels could correspond to tech levels, or, as soon as AI is possible all possible levels could be available. In the latter case, "dumber" AIs represent AIs with more constraints.
Constraints are what keep the AI from going SkyNet on the player. Higher levels of AI would produce higher risks of singularity even as they provide more productivity. Chance of risk might be weighed each year or even each turn (with higher AIs).
Consequences of singularity would be driven by a table with a range of consequences. Most likely: AI burnout / AI suicide--loss of all census/productivity. Less likely -- AI tumor (AI dumbly works to convert all system resources into some random build order and/or AI census). Worst case -- AI effectively becomes a new CM controlled faction with randomly generated AIX traits; roll on table for disposition towards parent race / life in general / or unique goals that AI faction will work towards.
Singularity might be mitigated with player investment in some version of "reaction control rods" that could shut down an AI headed toward singularity (better tech = better chance) with the associated cost of a turn or more's lost productivity.
I would like to see a balanced rule that challenges players to carefully weigh their greed for resources against the risk of creating Cylons.
Thanks for any/all thoughts on design/implementation.
Cheers,
Josh