It’s ironic that a referendum to change the voting system for people gets voted down by the system for ballot measures. Another example of why elections are bogus.
Hi Terry, my name is Adrian Zidaritz and my publication is also about redesigning democracy. In 2022, I founded a nonprofit called "Stronger Democracy through Artificial Intelligence". I bring it up here because the name of the org tells you how I think about doing that redesign.
As outlined in your article, the use of sortition brings the following to mind. (If you have the time to read the Reader's Guide for my publication,) I draw a system of democratic governance that contains an AI foundation model at the top, which model is trained through citizen participation. Various ways of tweaking democracy can be accomplished by varying the parameters of this foundation model.
Sortition would be an awesome use case. One may decide to use the ancient Athenian lottery process or more modern scientific stratified sampling and then "measure" the effectiveness of governance that would result from implementing sortition through one particular method.
That's a mouthful, I know. But I wonder if you would be interested in helping each other with this common goal that we both have.
I am interested to hear more about your proposal. It's always good to meet allies in this work!
I have two initial thoughts...
1. Measuring the "effectiveness" of governance is difficult because there is no universally agreed upon standard. Who should decide on the society's goals/values by which we measure efficiency? Philosophers, kings, experts, etc. or society in general?
If we believe that society in general should decide, then the best that empirical science has come up with for understanding such a large/complex system seems to be random sampling. My trust in sortition is based on my belief that people in general should control society and on the empirical evidence that exists regarding the usefulness of random sampling for understanding complex systems.
2. I worry about including AI in governance processes because, at least to my understanding, LLMs are not fully knowable by individual humans. Their behavior cannot be reduced to understandable mechanics and so cannot be effectively audited or held to a standard of "predictable and fair." So, with my current understanding of AI, I would not trust its involvement in distilling the goals/values of a society. However, I do think AI could possibly be trusted in the capacity of EXECUTING policy (not deciding policy) because its actions could be audited against the goals/values/standard of policy set by the society (via sortition).
I forgot the link. The Reader's Guide is reachable through the top menu of the home page. Please feel free to comment on any post that may draw your attention. Thank you.
It’s ironic that a referendum to change the voting system for people gets voted down by the system for ballot measures. Another example of why elections are bogus.
Hi Terry, my name is Adrian Zidaritz and my publication is also about redesigning democracy. In 2022, I founded a nonprofit called "Stronger Democracy through Artificial Intelligence". I bring it up here because the name of the org tells you how I think about doing that redesign.
As outlined in your article, the use of sortition brings the following to mind. (If you have the time to read the Reader's Guide for my publication,) I draw a system of democratic governance that contains an AI foundation model at the top, which model is trained through citizen participation. Various ways of tweaking democracy can be accomplished by varying the parameters of this foundation model.
Sortition would be an awesome use case. One may decide to use the ancient Athenian lottery process or more modern scientific stratified sampling and then "measure" the effectiveness of governance that would result from implementing sortition through one particular method.
That's a mouthful, I know. But I wonder if you would be interested in helping each other with this common goal that we both have.
I am interested to hear more about your proposal. It's always good to meet allies in this work!
I have two initial thoughts...
1. Measuring the "effectiveness" of governance is difficult because there is no universally agreed upon standard. Who should decide on the society's goals/values by which we measure efficiency? Philosophers, kings, experts, etc. or society in general?
If we believe that society in general should decide, then the best that empirical science has come up with for understanding such a large/complex system seems to be random sampling. My trust in sortition is based on my belief that people in general should control society and on the empirical evidence that exists regarding the usefulness of random sampling for understanding complex systems.
2. I worry about including AI in governance processes because, at least to my understanding, LLMs are not fully knowable by individual humans. Their behavior cannot be reduced to understandable mechanics and so cannot be effectively audited or held to a standard of "predictable and fair." So, with my current understanding of AI, I would not trust its involvement in distilling the goals/values of a society. However, I do think AI could possibly be trusted in the capacity of EXECUTING policy (not deciding policy) because its actions could be audited against the goals/values/standard of policy set by the society (via sortition).
I forgot the link. The Reader's Guide is reachable through the top menu of the home page. Please feel free to comment on any post that may draw your attention. Thank you.
https://adrianzidaritz.substack.com
super interesting to hear how this citizens assembly came about