Ask HN: Is anyone using agents to test a better version of democracy?

3 pointsposted 8 hours ago
by amelius

Item id: 50171447

14 Comments

mindcrime

7 hours ago

Not exactly, but I did start chatting with ChatGPT a couple of days ago, about creating simulations of large scale social organization with NetLogo. It was all pretty speculative / early-stage, as I mostly just wanted to see if the AI could create NetLogo (or something similar) at all. I am intrigued by the idea of creating simulations to try to suss out some possibilities around what a future economy without much (if any??) need for human labor might look like.

Disclaimer: I am not an economist, or a sociologist, etc. So this will be, at most, some random guy playing around with models with some help from an AI. And that's even if I find time to actually put into it.

silverFork

8 hours ago

Is somebody thinking to replace members of US congress each with an Ai agent that works for each region? Interesting idea, at least they can't be bribed and they don't seek to become lobbyists. They will also have very good memory and they can find a median of demands from them and they may be able to detect astroturfing. How can you test it out side the frame work of real population though to see how and if it can be subverted?

verdverm

7 hours ago

A citizens' assembly is an interesting idea I've heard proposed a few times now. The gist is that people are randomly selected, like jury duty, to participate in the legislative process, either crafting policy recommendations or as a wholesale alternative representation concept.

silverFork

4 hours ago

Random people is a good idea buy they the will interview them to filter them and disqualify people based on ideology, left, right, up and down. Maybe an Ai agent with transparent goals would be better. Some small country will try it for sure and it will work there but not in a massive and highly divided country in my opinion because the trust factor is a required element .

verdverm

4 hours ago

The politicians don't get to filter like lawyers do for jury duty

Ai is never the answer for human relations or societal coordintion

solyin

2 hours ago

Why not? What's even the question to begin with? If a human is having trouble relating to another human, but society is bereft of human support to address that issue in an affordable manner, why should AI not be used? If AI is used to build a website coordinating some aspect of society like a community messageboard saying the pool is closed until the filter can be fixed, all of society must reject it?

verdverm

an hour ago

In your example, Ai is not doing the coordination or allocation of resources. It was used to build software to enable that. Different things

I would put limits on Ai companions and Ai decisions over human lives.

silverFork

4 hours ago

Have you seen them select judges. They do filter for mundane and then select the crazies that typically hate women.

With Ai it will be about trust and whether it can optimize between wishes of masses, rights and work towards the best outcome.

Do I trust Ai? not off the box but if its pattern of past decisions are exposed, maybe. I don't hate Ai yet. I think people will corrupt them in time as they always do in future.

verdverm

3 hours ago

I'm saying that the law would not have politicians picking or filtering the citizens assembly. They would have no say in the matter.

solyin

2 hours ago

I made a book about it, dunno if that counts tho.

sjdh76

6 hours ago

Ironically moving fastest is China, UAE and Singapore :) You can google for their agentic initiatives.

verdverm

7 hours ago

No, now that I hardly write code, I'm reading and writing more human. We'll need to speak with each other, face-to-face, not through social media or agent intermediaries - if we are to build a better future for all.

Working on an essay for this subject, here is my current reference section

---

## References

#### Michael Sandel

- [The Tyranny of Merit - Book](https://www.penguin.co.uk/books/313112/the-tyranny-of-merit-...)

- [Why 'Merit' Should Be Rejected](https://www.youtube.com/watch?v=Vp5qsEuUOWo)

- [Democracy in Peril: Paths to Civic Renewal](https://www.youtube.com/watch?v=Qi8icKc9kKw)

- [Are There Things Money Shouldn't Be Able To Buy?](https://www.youtube.com/watch?v=zMg9Gjz8PKs)

- [Who Merits to Govern Us?](https://www.youtube.com/watch?v=-0UhSo5ZNX8)

- [Does the Winner Truly Deserve Their Success?](https://www.youtube.com/watch?v=03i_Ojbx5do)

#### Daron Acemoglu

- [What Happened To Liberal Democracy - Book](https://www.penguinrandomhouse.com/books/815318/what-happene...)

- [Building Pro-Worker Artificial Intelligence - Paper](https://www.nber.org/papers/w34854)

- [Conversations with Tyler](https://www.youtube.com/watch?v=e5O1Y-5phEU)

- [Foreign Affairs Interview](https://www.youtube.com/watch?v=I3iK7j36tI4)

- [Nobel Prize 2024 Lecture](https://www.youtube.com/watch?v=RmVL30sS2yU) ([All Laureate Lectures](https://www.youtube.com/watch?v=YcuxbYUW8R8))

- [Why Nations Fail: The Origins of Power, Prosperity, and Poverty](https://www.youtube.com/watch?v=A4eVoOlIuis)

#### Friends Together

- [Celebrating Daron's Book](https://www.youtube.com/watch?v=bKdfNztHq1k)

- [Daron interviews Michael](https://www.youtube.com/watch?v=0ZpINmi6C3o)

#### Bonus

- [Opus 23 - Ai does to us what American Cheese did to food](https://www.youtube.com/watch?v=1CbmC2aWhjY)

- [The Students - What the Greeks Would Call Our Government](https://www.youtube.com/watch?v=205IdzGJMJs)

- [Simon Johnson - 2024 Nobel Laureate Lecture in Economics (a good history lesson)](https://www.youtube.com/watch?v=iBwuEt3DqoI)

- [Michael Young - The Rise of Meritocracy](https://archive.org/details/riseofmeritocrac0000youn)

- [EFF Thoughts on Ai Issues](https://www.eff.org/issues/ai)