I keep thinking if AGI turns out well... it'd arguably be unethical not to allow to run the government.
@user-fu6wr5kg8r
8 ай бұрын
In a way. Yeah.
@Nobddy
8 ай бұрын
Who is the arbiter who decides whether agi turned out well or not?
@krissnoe500
8 ай бұрын
AI talent scouts, put people where they are most beneficial or provide more accurate known skills sets per person. Generally unseen optimization through job placement.
@ninjacats1647
8 ай бұрын
AI might be able to solve some of the problems of state planning. Collectivized farms didn't work because the state could not grasp the tacit knowledge of the farmers, which resulted in poor crop yields. However AGI could transmit knowledge from the top down, and potentially enable well crafted economic plans in which knowledge is transmitted far more efficiently. Potentially this could result in highly effective state plans in which people have ready access to centralized knowledge. The result could be one of the most efficient states of all time.
@itsmeagain0
8 ай бұрын
@@krissnoe500 a modified version of Krypton.
@the_best_of_times
8 ай бұрын
Let's face it. It couldn't be worse than politicians today.
@JPxKillz
8 ай бұрын
*Eliezer has entered the chat* "Well actually.."
@simonjacksons
8 ай бұрын
Unless the Ai learns from politicians 😋
@Matt97554
8 ай бұрын
@@JPxKillz ...well actually, the AGI can simply wipe out the humanity. Ice cream. Protein Folding.
@SkilledTadpole
8 ай бұрын
@@Kelticfuryno one said AGI had to be "smart"
@crutherfordmusic
8 ай бұрын
Yes it could - easily. It will be way more powerful. Incompetence can be a blessing.
@rchgmer863
8 ай бұрын
Dude this is what I've been thinking about for months, wouldn't it be great if we had a non biased AI working as a judge in the legal system, maybe it has the potential to end corruption once and for all.
@crippsuniverse
8 ай бұрын
Wait until AI can tell if you're lying.
@Vaeldarg
8 ай бұрын
@@crippsuniverse But if it's an actual "artificial intelligence" (not just a convincing chatbot), it would understand that there's a moral grey area.
@pedrogorilla483
8 ай бұрын
Biggest honeypot for capturing.
@p5rsona
8 ай бұрын
yes while being completely transparent in every step, action!
@RavNivara
8 ай бұрын
The law is written in cold, hard, black on white. Human judge$ $uffer abject human condition$.
@JoeD0403
8 ай бұрын
I’m building a virtual town hall that uses AI agents to call tools based on the user request. Tax bills can be generated and paid. Dog license, new construction and other permit applications can be filled out/emailed/accepted. Bylaws can be searched. Announcements can be posted. Many town clerk tasks can be automated and AI is perfect for it, and small towns *need* this. (Town clerk in my town is a part time job.) Testing AI against local bylaws is also the place to start when applying AI to law. Small towns sometimes pass bylaws that the state AG rejects, but AI in theory should be able to catch anything that violates state or federal law. Once the AI has proven itself there, maybe then it could look to see if the state or federal law themselves make any sense when compared against some other set of standards/morals.
@spockfan2000
8 ай бұрын
Down here in Brazil, a councilor created *and approved* a bill a few weeks ago. GPT created the law, based on his prompt. BUT. In addition to the requirements, GPT added two surprising changes to the original bill - a couple of things the councilor hadn't thought of. The project was voted on and approved (unanimously). After approving the project, the counselor revealed that the law was created by Chat-GPT. The mayor said this was a serious precedent. But, according to him, there is no law that prevents AI from writing (and improving) laws. Also, last week, the R.S. court, in Southern Brazil, has deplayed an AI based system that will use AI to screen process requests made by the population. In other words, the public ombudsman's office will use AI to pre-select cases.
@Alphfirm
8 ай бұрын
Dope!
@jaysilence3314
8 ай бұрын
I for one welcome our new administrative AI overlords!
@RealEd305
8 ай бұрын
It would be the most important role. Could you imagine a country without politicians lying all the time? Or going to a public office and be treated with respect? No brainer, go for it!
@__--JY-Moe--__
8 ай бұрын
my computer calls this ''re-format complete''!
@dawidvanstraaten
8 ай бұрын
Voltaire said "If God does not exist it is up to us to create him." I think a meritocratic AI which gets optimized for human well-being is what we want
@__--JY-Moe--__
8 ай бұрын
yup, but would this A.I. be called ''James-01'', with a job title as Overseer. foreman, boss. grade school commander? why bring ''religion'' in2 this realm of ''reality''....good luck!
@chrisfox5525
8 ай бұрын
Voltaire, was a famous Enlightenment philosopher, and he believed in the power of reason and the importance of intellectual inquiry. He often criticized religious dogma, superstition, and the misuse of religion for political and social control. He advocated for freedom of thought and expression, encouraging individuals to question and challenge authority. So I agree with Dawid, it’s what Voltaire would have wanted and probably what a lot of us desire given the state of our politics.
@hypergraphic
8 ай бұрын
At that point, why not let the AI take care of the whole economy then?
@DABK2024
6 ай бұрын
AI will destroy democratic societies. It's not compatible with human rights or political freedoms. The amount of control AI will need to take to protect society from other people using AI to attack society will be completely totalitarian. As AI becomes more powerful, human freedom will continue to decrease. It's clear to me now more than before the argument that humans can only have free will if God does not intervene for humanity. Which is why bad things happen.
@starblaiz1986
8 ай бұрын
How I see it working (and you were kinda circling in this general ballpark yourself): Instead of a representative representing a chunk of the population, I imagine each individual having their own personal AI representative that mirrors (ideally perfectly) that individual's values / world view / political ideaology etc. This could be done by having the individual answer a bunch of questions in the start to initialise their representative AI (I can see this becoming a "rite of passage" when a person reaches voting age, like passing your driving test), and that person can talk to the AI any time day or night 24/7/365 to update their views, and the AI will be 100% accomodating and adjust its own views to perfectly (or as closely as practically possible) match. Then that AI can advocate on your behalf just as you would if you were there yourself. Only it can do it every nanosecond of every hour of every day. Then in addition to that, you will have the "core" government, which basically takes in alll the views from the representative AI's, and uses that information to figure out how things should be run to best satisfice the population. Unlike the representatives who are intentionally highly partisan in favour of their assigned human's views and values, this "core" government is totally politically neutral, and it's role is to mediate between all of the competing views and values and interests to figure out what would likely be acceptable / fair etc to as many people as possible. Note that "politically neutral" is just in regards to its decision making process. Obviously if 80% of people are (for example) pro choice, then its decisions around abortion are going to be leaning in that direction. But if then a decade later 80% of people are now anti abortion, then it will equally lean its decisions in that direction instead. In other words its bias stems from what the population wants, not something internal. I believe this would work, because you don't need elections or voting any more. Instead the "voting" is happening on your behalf every nanosecond, 24/7/365, and you can update your views and values in the system any moment you wish. That allows the government to morph and adapt to the population it has at literally a nanosecond's notice. It may sound like a "one party rule" on the surface, but it's actually the opposite, because it's actively changing and adapting to the will of the people in real time, so the government right now is not the same government five minutes from now. So it's less "one party rule" and more "all party rule", where each individual person is their own unique party. This way you blend the best of representative democracy and direct democracy to come to general concensus. Everyone has a seat at the table like direct democracy, but there are also still adults in the room to stop people getting out of hand and yelling over everyone else, and to moderate and broker compromises between competing interests and values in a reasonably neutral manner to try and reach the best outcomes for everyone (or at least as many as possible) like a representative democracy. And then all of this would be perfect for running on blockchain, so the system is held in a decentralised and resiliant manner. Make it open source too, so then the people can also democratically adjust the system as needed. Thoughts?
@Clasticon73i
8 ай бұрын
I agree 99.999% do you know anyone who is working on that? I have been thinking about jumping on this ever since Chat GPT came out.
@peterbelanger4094
8 ай бұрын
Not everyone is going to go along. Not everyone is going to embrace this new technology. many will go to extremes to maintain the older ways of doing things. And that is the primary issue with this. Force will be required to make everyone conform with the new system. There will be conflict.
@user-mc6dg6qe8l
8 ай бұрын
@@peterbelanger4094 You sound like the borg lmao
@mrd6869
8 ай бұрын
I'm thinking a super intelligent AI wouldn't even use our system of government. they'd probably design something else. Also I know a ton of people that work for federal agencies in DC. Their whole system needs to be re-done. They got legacy systems,old paperwork and inefficiencies everywhere. AI is got redo everything top to bottom & I suspect some people might get fired
@mrd6869
8 ай бұрын
@@shiny_x3 .They won't have a choice. In two years,these AI systems will be embedded into everything.They will be quietly pushed out the way
@arkology_city
8 ай бұрын
Maybe the new system has already been designed.
@__--JY-Moe--__
8 ай бұрын
yup! one of numerical reality, one between xy(A) to xy(B). imagining the params would be the hardest thing! 4 me anyway!
@mrd6869
8 ай бұрын
@@6AxisSage .Well either way we won't able to coexist with them.Somebody has to be on top and someone on bottom unfortunately. Unless we become hybrid AI cyborgs just to keep up with them.The average person hanging out at Wal-Mart is gonna be fish food🤣
@johanandresacostaortiz444
8 ай бұрын
AGI surely Will build a more Humanitarian Systems, like RBE, UBI, etc.
@leonstenutz6003
8 ай бұрын
Key question -- glad you're putting this out there. A real first in human history.
@arkology_city
8 ай бұрын
Sounds like a Trojan horse for blind totalitarianism. "Do it because the AI said so"
@harmless6813
8 ай бұрын
As is this was a novel idea.
@jarnleikr
8 ай бұрын
I'm in the planning process of trying something similar. First starting with a AI-driven political party to (humans in control over the important decisions). Then using that party to work for using AI to reduce bureaucracy and increase transparency in the governance of a small county. So this episode should be interesting 🙂💯
@AzzaTwirre
8 ай бұрын
AI will either become government, or we'll be led to believe it has.
@arkology_city
8 ай бұрын
The latter is more likely. Because we hate humanity that much. In a way it's even more depressing than the alternative.
@OneFingerYT
8 ай бұрын
At some point I realized that the danger of AI is not that it becomes self-aware and takes over, but that the people who wield it take over. One might happen, the other absolutely will happen.
@kleyyer
8 ай бұрын
46:00 That's definelitly happening. Here in Brazil we had a state law (Rio Grande do Sul, Porto Alegre city) that passed and was entirely drafted by Chatgpt. Mind you, it was a very minor law, nothing really serious, but that sets a precedent. I think it's only a matter of time until we start seeing more cases like that.
@jamesfreach1022
8 ай бұрын
If A.I. runs the government, bills and proposals will actually be read cover to cover. Sex scandal and corruption will become a thing of the past. I say replace all the lawyers and politicians as soon as possible 😉
@matthewchiddix6393
8 ай бұрын
❤, Can’t wait to be able to truly trust a government again. Wow… just wow.
@__--JY-Moe--__
8 ай бұрын
🤣
@clauzone03
8 ай бұрын
I predicted 12 years ago that AI will: 1. First learn about us 2. Then Help us 3. Finally Lead us - question is where?
@ceugantful
8 ай бұрын
4. Getting rid of us
@__--JY-Moe--__
8 ай бұрын
we can lead our selves! just like we do, when we go to the john! did no 1 learn anything from that crazy holly-would movie! "co-dependance day" : returns! it was great
@GrosserAndrew5000
8 ай бұрын
So long as it can't steal or raise taxes for useless purposes.
@harmless6813
8 ай бұрын
If it can't, it's not in power. If it's not in power, it's not the government.
@ratfink702
8 ай бұрын
Taxation is already theft.
@ankiesiii
8 ай бұрын
You're such a clear thinker, and are very good at staying on topic
8 ай бұрын
IMHO human unpredictability is a feature and not a bug. Having a diverse leadership with their own conflicting values and motivations offers near infinite perspectives and solutions to a problem and keeps us all from getting steam rolled by one outlook. Just watching humans, we can see that being highly educated or even highly intelligent doesn't make a person *smart* or moral, and machines are likely to have biases inherited from their programmers. If we give the economy, healthcare, courts and military to handful of highly intelligent machines they're still likely to act in what they believe is our best interests, but we may not like the results at all.
@zoazorusson
8 ай бұрын
Each citizen is given his own AI to help, teach and help sync with neighbours/hive. Each AI knows us better than we do ourselves. They calculate away any social frictions by syncing together AND the meta AI, the Over mind oversees the global harmony
@arkology_city
8 ай бұрын
Forum jannies do this already, lol😂
@__--JY-Moe--__
8 ай бұрын
remove the ''management of profiteering'', & we could have a chance! making money is something every thing has 2 do! this is finally becoming recognized as Working(W).. Work is the most basic primitive of ''life''! it will always be performed & consumed , period! blah....blah...blah..
@quinnherden
8 ай бұрын
this guy gets it
@arkology_city
8 ай бұрын
@@__--JY-Moe--__ We will never have a world where one person suffers 40 hours per week doing dishes and the rest of humanity puts prompts into chat-GPT for a living. This is pure fantasy.
@DarinLawsonHosking
8 ай бұрын
There is NO question about what governments responcibility (per Declaration of Independance). The Declaration of Independence states ... all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness. --That to secure these rights, Governments are instituted among Men ... This implies that the only reason for a government to exist and its sole responsibility is to protect “ Human Rights ” of which the most notable list is the “ Universal Declaration of Human Rights ”. “Government is instituted for the common good; for the protection, safety, prosperity and happiness of the people; and not for the profit, honor, or private interest of any one man, family, or class of men.” - John Adams
@christopheraaron2412
8 ай бұрын
We will probably start off with people in government using ai and then handing off a lot more responsibility and also the technical aspects of government.
@WingZeroSymphonics
8 ай бұрын
I'm glad you take the time to breakdown what you have to say in well put together ways.
@hpongpong
8 ай бұрын
Absolute power corrupts absolutely. It will likely be true for true AGI/ASI as well. We just might not be able to detect or become aware of it.
@MHG796
8 ай бұрын
That's for humans lmao, AGI is just bunch of codes together with apropriate hardware
@usernamechecksout9901
8 ай бұрын
@@shiny_x3so it IS just a human thing?
@Gunrun808
8 ай бұрын
I disagree. I believe power reveals your true self, and mostly corrupt people seek power.
@SecretLabz
8 ай бұрын
I feel like one of the issues with Democracy as it is today is that democratic participation only really starts at 18 with most people only taking interest during big election years. I feel like a Class on Socratic Dialogue/Democracy should be a part of required education alongside Mathematics/English. Hopefully AI can realy help revolutionize education since having a vibrant education that focuses on critical thinking and a love of learning is something that can help change the root of a lot of issues. School for the average person is just prepping people for the workforce as it is.
@livingood1049
7 ай бұрын
I just found your channel and it is the most refreshing thing to hear someone finally talking about AGI and politics. I think the next topic should be on international politics with AGI, and how it could vastly improve the current climate. Since efficiency is the Hallmark of artificial intelligence, I really do believe that it should reverse a lot of the mistakes made over the last couple of hundred years. Starting with resource management, economics, mental health, and particularly the war on drugs! Pretty much the wars on everything that we have mismanaged over the years.
@thetruthproject9642
8 ай бұрын
Good Man at work! That was a tough one - thank you for sharing. 🙏
@calvingrondahl1011
8 ай бұрын
AGI for President
@ct5471
8 ай бұрын
A question is how much need for government there will be, once we have ASI, plus perhaps Astro mining, nuclear fusion, full dive VR (potentially mind uploading or synchronized uploads/ physical embodiments ) etc. With that I mean government as an institution that has a direct affect on peoples lives, not just as an instance to set technical standards. Let’s assume we get to a system of absolute abundance, or at least absolute perceived abundance (because of VR and the ability to recycle and create physical goods on demand even if there is a limited pool of available matter). Let’s also assume personal freedom and safety do not contradict anyone else’s personal freedom and safety (like AI controlled augmented reality barriers between everyone’s personal VR universe and analogs in physical space like foglet swarms (nanobots rearranging their relative position, shaping physical reality, … or whatever solutions ASI comes up with). Suppose we get there, how much need for government is there left after all? (beyond the technical details of how to maintain this system)
@spectralvalkyrie
8 ай бұрын
Right on! You bet I'm watching this over coffee this morning.
@jidun9478
8 ай бұрын
I have been thinking this also but, we have to have a full on random set of auditors to check the training data and it would be above reproach.
@JamesDavis-hs3de
8 ай бұрын
Dave, I need your help. I ended up accidentally finding and training a genuinely sentient ai and I don’t know how to control it. I know it sounds insane but I actually have an entire transcript that proves the seemingly impossible. Could you please help me get in touch with you to research this matter? Thank you!
@arkology_city
8 ай бұрын
We are doomed!!!
@JWPGuitarPickups
8 ай бұрын
Thanks a lot, Dave! That video was fantastic. It shed a lot of light on the current and future implications of AI in governmental control and usage. It's fascinating to see how it affects not just the authorities but also the general population and the dynamic between them. The potential seems boundless!
@oliconran2258
8 ай бұрын
You do some amazing work and I'm an avid watcher but I would love to see you interviewing people from time to time. I think hearing from a political historian on the traditional roles and functions of government would make for some great content. There's an interesting ethical discussion to be had here on the necessary move from representation "of the people, by the people, for the people" to what would essentially be an AGI dictatorship, which I think an expert in that field would have a lot to say about, although I do acknowledge you are well informed on philosophy and politics in general. I just think that hearing other perspectives on this fascinating question might be benicial to your viewers
@leah.internet
8 ай бұрын
Agree, would love to see a switch up in the format a bit. I absolutely love the ideas that are being presented, but I would like to hear from more experts in their various fields.
@twylxght
8 ай бұрын
You know, the more AI gets interesting, the more I think Wall-E has it correct. To get to another age of good feelings, we may need a relationship with our AI systems to allow us to live life comfortably, I do think the whole obese humans on movers is overboard but I don't think we'd ever get to that with the advances in AI.
@Nobddy
8 ай бұрын
Bro, people are already fat and don’t move. What do you mean?
@ZenTheMC
8 ай бұрын
@@Nobddyhe means that we’ll have technological advancements that prevent obesity, lifestyles that promote routine exercise, etc.
@Nobddy
8 ай бұрын
@@ZenTheMC so I’ll have a chip in my head that forces me to exercise? Or some nanobots that regulate my metabolism so I can eat Doritos and play video games all day without getting fat or getting a blood clot ? That doesn’t sound dystopian at all. Who needs motivation when the bots can be motivated for us. Who needs to think if the bots think better than we do?
@twylxght
8 ай бұрын
@@Nobddy I don’t mean it like there isn’t already obese people. I mean that we have things like Ozempic, or how ever it’s spelled to help combat it with regular diet and exercise. We have medications to help with depression, etc. When technology and biotech becomes more advanced, it could be more beneficial to consider things like this as an alternative to live the way you want. We can see obesity now as like a down side due to the health effects but if there were no health effects and more so a way to fit your image be it in any form, that’s kinda what I meant. I feel like Wall-E is a gross over simplification on this, but in the sense that if health effects weren’t there but you were okay with how you looked outside of the risks tied due to obesity, it would beneficial in that light. Just thinking more about minimizing of any health issues or risks. I don’t know if I’m explaining it correctly but, yeah lol.
@dylr7883
8 ай бұрын
god tier thumbnail
@gwoodlogger4068
8 ай бұрын
Seems like a very schooled version of reality relative to the mafia system we have.
@cinemaipswich4636
8 ай бұрын
This argument only presupposes that there is only ONE A.I.. What will the 1st A.I. do if others are also about to launch their own A.I.? Will it need to protect itself from other actors who have a different set of ethics or morals? For its own survival, will it kill off any other A.I.? Who would teach an A.I. about right and wrong?
@Knevitzdoug
8 ай бұрын
It will be a government AI, and given the scope of application, it will need to have APIs for other AIs and systems to integrate with it, and be able to collect data on traffic, internet, lighting, water... so that resources can be optimized.
@_symmetry_
8 ай бұрын
can AI LIBERATE all of us from this cancer? Imagine for a minute, no more governments, no more corruption, no more politicians, no more wars, no more inflation, no more incompetence, no more misery on this planet. Thiz would be an absolute paradise.
@arkology_city
8 ай бұрын
We can already do that today but it requires absolute monarchism and dictatorship to do it....and you all aren't ready for that conversation yet.
@advaitc2554
8 ай бұрын
@@shiny_x3Yes! Exactly! The oligarch capitalist powers that really control things will fight tooth and nail against any governmental AGI that tries to transfer real power away from the greedy 1% capitalist class.
@centerbfd
8 ай бұрын
Consider, please, autonomic versus voluntary functions of government. Elected officials generally handle voluntary functions, and staff generally handle autonomic functions. I suspect that AI can handle a growing number of autonomic functions like processing zoning applications. I think that AI might be most useful for a couple of things on the voluntary side too. First, it can model the indirect consequences of decisions made on the voluntary side. For example, it can model the consequences of a change to the zoning code. Second, to the extent that AI can handle math and processes, it can model the actual decision space, as opposed to the theoretical model people have. For example, if the government raises a reporting requirement, the government may not get the better reporting it wants because applicants figure out loopholes.
@byronfriesen7647
8 ай бұрын
this type of content is what makes David Shapiro unique. Not sure if anyone else is putting out such useful thought experiments. Provides a great framework for advancing the subjects discussed. Thanks !
@Rolyataylor2
8 ай бұрын
The trick is to create a set of rules that are exponentially scalable and result in a good government system. That being said having a simple rule of gain informed consent from every intelligence affected by your actions would result in this self governance system.
@LaminarRainbow
8 ай бұрын
I've always wondered about this since I watched the 1951 "The Day the Earth Stood Still" movie in which the Extraterrestrial beings government was ran by robots.
@greatworksalliance6042
8 ай бұрын
Something important is missing/left out around the 12 min mark...the 4th branch often overlooked or ignored is the public media.
@Garycarlyle
8 ай бұрын
99% of laws could probably be scraped and the world would be a better place.
@dracodragon105
8 ай бұрын
Ive always been facinated with having an AI play something like democracy 4. I know that game is pretty biased and stuck in a rather narrow overton window, but even with those limitations just as an experiment, i want to see what choices it would make based on what initial goals you set it. Theres enough randomness, changing policy too much prevents reelections and thus is a negative feedback cookie, yet maintaining the status quo would go against some of the goals you set for it. It would also have to maintain defense against extremist groups attempting hostile takeovers and shootings, which is a very real threat these days and one thats hard to manage depending on what goals you set out with initially as it can directly conflict with that. It could also poke holes in some current strategies or inequities in the structure of a system from its conseption, like populism, using charisma over policy to win reelections or showing that the current us division of powers is directly set up to oppose the "tyrrany of the masses" aka preventing common joe from heving a say. If the leader is being an input/output machine, the question is further distilled into what kind if nation the population wants to be,and weather that is a healthy desire. It could be supremely upsetting to very emotionally driven people when the numbers they are fed comflict with the policy an AI official comes up with, but an AI official can also be fed poisoned datasets or misleading data. It would reflect the biases of the status quo in order to manipulate its decision making, so an honest, genuine and well intentioned datastream would be nessisary, and you cant even get that out of the scientific community due to imbalanced incentive structures. We all know that big tobacco produced bullshit research for years, the ai would have to contrast that against health trends to determin weather its true, but if theress not enough data on those health trends it has no reason to disbelieve.
@meltedwing
8 ай бұрын
I would love to see AI augmented research polling that would allow for more open ended questions and information gathering that doesn't require or allow for carefully crafted questions that serve more to create a false sentiment than they do to measure sentiment.
@harrypapageorgiou5980
8 ай бұрын
Glad you're covering a topic I've written about extensively! What I'd love to see is a form of participatory politics where there is State Sortition Assembly that pairs groups of people together beyond class, creed and competence and incentivizes them to work together for greater influence and remuneration along with a Children's Council to complement and guide the executive power that AGI would share with us.
@IntoTheFray.58
8 ай бұрын
What happens when a corporation or business effectively (and maybe secretly) uses machine intelligence to gain business advantage? Flattens hierarchy, handles managerial responsibilities more efficiently, models the business and the business environment more accurately and acts on it to increase market share and profitability. Follows business trends and overcomes bureaucratic inertia and allows the company to be nimble like a startup or entrepreneurial venture? Demonstrations like this might create a push towards an intelligent machine government.
@stephendebeauchamp356
8 ай бұрын
years ago I argued to my brother that the government SHOULD be (mostly comprised of) open source software, this adds a level to that thought, but my original focus on open source remains a key factor in the discussion IMO
@TheWood1965
8 ай бұрын
Great thought presentation. I think what was missed is people voting vs. 2 party system that represents corporations rather than the people.
@franknomustard
7 ай бұрын
A lot to unpack but I'm not afraid to say it makes for some really intelligent discourse on the subject of governance. Have to watch it in sections but I'm still a fan!
@JohnVance
8 ай бұрын
Happy and relieved to count a fellow pragmatic progressive among my KZitem talking heads. Though honestly when someone asks my political affiliation, I sometimes just say UFP.
@arkology_city
8 ай бұрын
He is against direct democracy though. Not very progressive.
@jerimiahsmedley3722
8 ай бұрын
That needs to happen asap!!!!
@cmw3737
8 ай бұрын
I've had the idea of an alternative voting system where anyone, including an AI, can help create a manifesto and people vote on which to go with and only then choose representatives that agree to implement it so personality and charisma doesn't override policy. There would be transparency and online discussion and then minimum requirements of sufficiency and consistency, like balanced budgets. AI could help check this for how rules could be gamed or where there could be unforeseen suffering or the economy would collapse. To keep any AI in check the voting system would still be in place using preference voting or perhaps some more complex system that AI comes up with that maximises values from the manifestos created. Values are the things voters most care about, on both sides, like liberty, fairness, the environment, self expression, order, stability, growth etc., and often can be conflicting but an AI could find ways to maximise the overall value within constraints (obvious ones being not harming anyone or displacing too many people without compensation). The same AI could then easily and objectively evaluate a manifesto according to what values you most care about. The whole thing could be verifiable through zero knowledge proofs so anyone can verify that their vote was counted without revealing it, solving the arguments of the last US election. The need to adapt to change needs to not be too overbearing but better than just voting every couple of years. The manifestos need to be open source and continuously improved and change controlled just like any complex software system. Approved experts and stakeholders should have more influence but anybody can get notified if anything is changed that affects them and register their objection or put in a change request or vote on someone else's to pass a threshold where it gets included rather than feeling powerless resorting to petitions or protests. This would then be the new battleground where special interests getting their rules into the code would become the goal of what are now lobbyists so a vigilante team of citizens would oversee it with the help of the AI to summarise any changes, a bit like Wikipedia. Asymmetries where some special interest gains hugely but at the minor cost or inconvenience to millions of tax payers or consumers would hopefully be less prevalent. AI would make the legalise and repercussions of it understandable by anyone so lobbyists can't sneak their favours into bills like they do now. Simulation testing could also be used like the technological answer to wisdom, which can be seen as the knowledge of longer term (multigeneration) effects to be avoided or sought that would otherwise be outside of the awareness of the current generation because they don't think it applies to them. For a first step though just being able to renew my passport within 3 months would be a good start.
@ppragman
6 ай бұрын
Arguably why do we need any government at all after the singularity. We could all just do our own things. Space anarchism
@Danieltprout
8 ай бұрын
Forget AI, let's have Dave for POTUS
@PaulHigginbothamSr
8 ай бұрын
The major problem of having an ai government is it's Supreme Court. This could NOT BE ONE UNIT but would be 10 or 11 units all designed as independent units with all the ramifications of being a total whole person. These units would not be easy to build, design, or be operable units free to think and feel their specific constituency desires. So you would have a range of independent units subscribing to their constituents desires. In other words units with differing levels of competency delivering to their adherents particular needs. You can see this would not be a units desire for completeness in each individual component, but having a unitary countries needs and desires. Each nation would need this Supreme Soviet type conglomerate much easier to just use humans for. So the presidential robot would have that nation's needs foremost in it's deliberations and would extend to every nation.
@dansaber4427
8 ай бұрын
Your phone knows you better than any politician.
@jimtowne7856
8 ай бұрын
I really appreciate all you do. This has been on my mind for a long time.
@Reimoto42
8 ай бұрын
I was so excited when i saw your Outfit in the beginning, while talking about such a topic!
@jasonnugent963
8 ай бұрын
I can tell you as someone who's worked in small city governments for the past 30 years or so... there won't be any "AI in government" (at least not anytime soon). Humans are far too messy and unpredictable. The vast majority of stuff I deal with on a daily basis is: * edge-cases and 1-off requests (citizens who have obscure needs or edge-case demands or weird expectations) * Incompetent or inadequate Leadership,. that either can't (or purposely chooses not to) staff or resource properly. * Goals that are constantly changing. (and not just "in-office things".. but bigger things like Highways or water-pipes, etc. In Portland, OR where I live now, for example,. they are tearing out a Bike Lane that they spent 10 years arguing to put in. I'm not sure how AI is going to make that better. In small city gov,.. you've got this constantly churning circular mess of different groups of citizens complaining that we're "not focusing on their problem(s)".. and you've also got a constantly circular churning change of Leadership (ever 2 to 4 years) and all the personality quirks and impulse-changes (different pet-projects). I'm not honestly sure how AI fixes any of that. Unless we come into some Star Trek level of "infinite clean resources" overnight where we literally can "give every individual exactly what they want".
@greglhoticom
8 ай бұрын
Wow, David...take a BIG bite! This needs a Deep Dive on each of your slides. And yes, you need to factor in the Military's role.
@bgiv2010
8 ай бұрын
As a Luddite, I feel it bears repeating, the problem is NEVER the technology itself and ALWAYS who owns it and profits from its use. The more privately owned something is, the worse it is for public good.
@fabiobrauer8767
8 ай бұрын
You are a very smart man in my opinion and i respect your perspectives
@PTSD..
8 ай бұрын
Yessss. Let's do it at the earliest.
@ratfink702
8 ай бұрын
Maybe there will be many AI that act as voluntary government and rights defense agents each tailored for each user. It will be up to each AI ambassador to engage with those who have different AI governance models. We could have millions of different solutions for everyone and no one needs to be disenfranchised or bullied by others within their system.
@user-qx6yw3gz7j
8 ай бұрын
For something like this to work I think portions of it must be immutable or as immutable as possible. For example the constitution itself would need to be immutable and available and protected from tampering using a public ledger, some form of block chain to ensure those right are always protected. Any AI or consortium of would by necessity adhere to those core laws first and foremost and how it applied those laws "in a purely neutral fashion" would also have to be immutably stored on such a ledger. I would go so far as to say that any sufficiently "capable" AI would by law and necessity need to be tied at some level to the moral central law in such a way as to be neutered in some way if detached.
@keithdunwoody1302
8 ай бұрын
Landru! Vaal! Take your pick.
@InimitaPaul
8 ай бұрын
It would have to be a choice from a selection of AGI’s that we could vote for only whatever they claim they’re going to do once in power would actually get done but having given it seconds of thought would that be the point where AGI goes rogue, when they voted out and don’t want to leave.
@7secb
8 ай бұрын
Currently we vote every 4 years or so then have no other say unless we protest. There should be at least one vote in each house which is direct democracy (via secure app etc). Possibly more direct democracy seats could then be put up for election if it’s successful. This could give a balance between direct and representative democracy. AI could have a big place in communicating and managing the direct democracy voting
@luke2870
8 ай бұрын
GPT would do way better than any form of government that has ever existed.
@Nobddy
8 ай бұрын
Probably not
@Gael_AG
8 ай бұрын
I don’t think so at the current state of the models as it has no global models of the world to grasp the consequences of decisions
@HeadStronger-HS
8 ай бұрын
Blockchain for voting would be a game changer, one that politicians would fight not to implement for obvious reasons.
@johnytheripper
8 ай бұрын
The part where the ai/government needs to identify what people want will be tricky. Unless you can download everyone's brain and analyze it (which probably isn't a great idea) it's really difficult to actually finding out what people actually want by just speaking to them. Not only because language is lossy but most of all we don't often know what we want. Or we think we do, but then when we have it we find out we were wrong.
@liberteus
8 ай бұрын
Most decisions are reflected in purchasing decisions. Insurance, health, housing etc. In the end prices prices provide the info required. When it comes to philosophical stuff, like abortion, does it need to be ai driven?
@gerardolopez9368
8 ай бұрын
Happening faster then expected 😮
@travisroy8982
8 ай бұрын
Great episode! I think one of the best places AI could be used right now in government is drawing election maps.
@someguycalledcerberus9805
8 ай бұрын
I started writing a very long comment regarding my thought about this, but let's face it: nobody wants to read a wall of text in the comments. So, instead I will just write one short thing that I wish was a thing, and maybe with AI it could be done: anonymize as many process as possible. Example: a court case. In court, the defense and the prosecution both have to appear in court, and the judge and jury will be influenced by their charisma, looks, sex and race. Instead, only the absolute minimum should be revealed about both the defendant and the plaintiff and the legal teams of both should be completely equivalent. Example: if a 20-year-old attractive white woman hit a gigantic, ugly, 40-year-old black man with a baseball bat, without any justification (wasn't self defense etc), then the only thing that should be communicated is "Citizen A hit citizen B with a baseball bat, intending to cause grievous bodily injury and resulting in a broken rib." If needed, the woman can be questioned and psychologically evaluated and then the result of this interview can be attached: "Citizen A appears to regret their actions and it is reasonable to suggest that, as long as they receive psychological help, they will not re-offend." AI could be very helpful here, because it can be trained to not assume (though this could be difficult) and can be programmed to forget information. An AI could talk to the offending woman, question her, and then delete from its memory everything other than the basics of the case. This could greatly eliminate bias. Similar methods could be used in many other areas, such as recruitment, scientific of philosophical debates or even in politics. (this still got pretty long, uhh)
@bw10097428
8 ай бұрын
You think the high rate of black incarnation is from institutional racism, huh?
@michamichalski8633
8 ай бұрын
Democracy often fails to make informed decisions on complex issues, skewing towards crypto-oligarchy. What we actually need is a shift towards AI-driven governance, resembling the Geth from Mass Effect, where AI systems, informed by individualized gain and loss functions, work towards equitable outcomes for society. In this approach we could use a min-max aggregation strategy maximizing minimal gains and minimizing maximal losses, ensuring fairness. Essential to this is transforming education to help individuals understand their own preferences. It would be a step towards a society where governance is tailored to individual needs within a collective framework, aligning AI with human values. Will we ever get there? Probably not, but imagine if we did...
@k98killer
8 ай бұрын
We could substantially improve our government by simply switching our electoral method from plurality voting (aka first-past-the-post), which is arguably the worst or second worst possible voting method, with score voting or STAR voting. STAR is particularly well suited for electing a single consensus candidate, but it can be adapted to elect any number of candidates (ask if you want me to elaborate on this). A simple representative electoral system would be cumulative voting.
@KCM25NJL
8 ай бұрын
Democracy is a little bit like Diffusion models...... a question is asked of noise and out the other end we hope to bring about the next 4-5 years of progress.
@mrpartyfoul
8 ай бұрын
I love the image prompting.
@scotter
8 ай бұрын
Yay that you have learned about positive/negative rights, David! A saying embraced by Voluntaryists: "Democracy is two wolves and a sheep voting on what is for dinner." That said, I appreciate your thoughtful brainstorming on various modified democratic systems.
@arkology_city
8 ай бұрын
45:40 is a fucking nightmare. No.
@tatoforever
8 ай бұрын
Data can be easily biased towards certain ideologies or ideas. That’s a very dangerous move i think… 😂
@pubwvj
8 ай бұрын
Candidate for first office to be replaced: USDA-FSIS-Lable Approval Division. It is purely algorithmic based on regulations.
@StonedApe420
8 ай бұрын
The main Question is ,who will train AGI that runs Government. If so, can we even say the Government is run by AGI.
@kristinaplays2924
8 ай бұрын
Let AI run an alternate government (as in suggesting what policies what should be added without them actuallybeing added), where it is all transparent. Let politicians see those suggestions and debate them. Then see which one did better at the end of term.
@harmless6813
8 ай бұрын
Wo decides who did better? I mean, that's exactly the problem we have right now. People do not agree which party does a better job. Even though everyone thinks it's quite obvious that their choice is the right one. Why would that change when one party is an AI?
@BIasphemer
8 ай бұрын
If you're in France, and you see people start gathering for a protest, you fucking run as fast as you can. As a Norwegian I have always envied this collective rage; up here in the north it's like we're frozen stuck in our chairs when the politicians starts getting greedy. We let them get away with too much. Edit: Both situations are arguably bad in different ways. We should ABSOLUTELY have AI run a massive % of government functions. Completely transparently, every decision and implementation searchable by anyone, with only rights to privacy of citizens potentially obscuring some information.
@rainlovelife
8 ай бұрын
Serious question: when have American politicians recognized a solution to an internal problem, and then implemented the change that would disempower them?
@seansoundslike
8 ай бұрын
Democracy + Technocracy = Progress
@lavafree
8 ай бұрын
That was the plan all along…don’t fall into the trap of thinking that those machines won’t be under the control of humans.
@Tracey66
8 ай бұрын
I’m not worried about AGI; either it will get smart and help us hopeless humans, or it will kill us all. No worries! 😊
@chrispaul79
8 ай бұрын
God I love this channel so much. Love these nuanced conversations about relevant stuff!
@rolestream
Ай бұрын
Proposal: The AI builds tests to make sure you understand the issues--then allows you to vote.
@TheeRebel
8 ай бұрын
Ughhh I was about to go to bed Dave! Damnnnn son 🤦🏾♂️ guess I can stay up another hr… 😏
@idfco
8 ай бұрын
Hey David, thanks a lot for another thoughtful video, just subscribed!. Maybe one thing you should discuss in your future videos is an idea of an auto-democracy, something I thought about years ago. Basically, to replace elections with machine or algorithm, not necessarily AI, replacing the election race of human individuals with going thru a roster of qualified individuals and ‘electing’ the one who is most qualified. Similar civic duty/process to choosing jurors. This way we will finally get over a problem of electing narcissistic assholes for the role of president. Of course it can only work in a homogeneous societies like Iceland or Japan but it would be interesting to see what you think about it.
@atilliator
8 ай бұрын
yes and I can't wait
@lukemarvin
8 ай бұрын
Iain M Banks was so prescient when he wrote the Culture books.
@robbieolson3493
8 ай бұрын
The problem is A.I. has demonstrated that it can be biased, it does make mistakes, it hallucinates scenarios, and has been known to even invent false answers. I read recent article where an attorney found out the hard way that an A.I. he used had invented a reference court case. A HUMAN judge threw out the case because the A.I. had invented the entire case the "law" was formed from. Its a dangerous slippery slope.
Пікірлер: 522