Two Minute Papers - 2020-12-19
❤️ Check out Perceptilabs and sign up for a free demo here: https://perceptilabs.com/papers 📝 The paper "The AI Economist: Improving Equality and Productivity with AI-Driven Tax Policies" is available here: https://blog.einstein.ai/the-ai-economist/ 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible: Aleksandr Mashrabov, Alex Haro, Alex Serban, Alex Paden, Andrew Melnychuk, Angelos Evripiotis, Benji Rabhan, Bruno Mikuš, Bryan Learn, Christian Ahlin, Eric Haddad, Eric Lau, Eric Martel, Gordon Child, Haris Husic, Jace O'Brien, Javier Bustamante, Joshua Goller, Lorin Atzberger, Lukas Biewald, Matthew Allen Fisher, Michael Albrecht, Nikhil Velpanur, Owen Campbell-Moore, Owen Skarpness, Ramsey Elbasheer, Robin Graham, Steef, Taras Bobrovytsky, Thomas Krcmar, Torsten Reil, Tybie Fitzhugh. If you wish to support the series, click here: https://www.patreon.com/TwoMinutePapers Károly Zsolnai-Fehér's links: Instagram: https://www.instagram.com/twominutepapers/ Twitter: https://twitter.com/twominutepapers Web: https://cg.tuwien.ac.at/~zsolnai/ #taxpolicy #taxes
"AI will replace the most hated dangerous jobs" -> AI replace politicians.
Sounds legit.
@イカの嵐 (Storm Squid) that would be great but i would assume there would a skew of votes due to the massive numerical differences in people of a particular social class so say theres a 4 to 1 difference between the lower and upper class most systems would be set in favor of the lower class which would lead to equity but eventually that would be surpassed and we’d fall out of equality again with most of americas propositions being set by the lower class
@8’ ...And goals that benefit the "lower class" are somehow... Inferior? Or something?
That will not solve any problems. Those who manipulate politicians for corrupt, power-hungry reasons, will just switch to finding ways to manipulate the people running the computer that runs the AI.
You make the false assumption that the politicians are at the top of the power pyramid.
Essentially, the AI will be just as controlled and corrupted as the politicians, except those in actual control will have a new mask to hide behind.
would have been funny if the AI started stealing from each others and avoiding taxes
you mean like skynet makes decisionz for us? umh nope.
Can we just appreciate the fact that they didn’t just simulate, they made graphics and animations too!
Imagine building a simulation for tax policy and spending hours with the water physics
@Bob I hope you realize that our brains are, basically, really really complicated "if-then" statements by that logic. I don't think you understand how intelligence works, let alone artificial intelligence. Then again I doubt you have experience with either.
Not to mention, true AI isn't just if-then statements, true AI can alter its own "instructions" in response to inputs and success of the outputs, which is the kind that's being used here. Look up "learning algorithms".
@Aidan Haynesworth AI does have bias. First of all, construe "AI" here to mean "trained AI", as untrained AI is relatively useless to society (which is the context in which we're discussing the merits of AI, given that this is a paper about economics). What is AI trained on? Data and/or environments. And who collects/curates that data, and designs those environments? Humans. Do humans have bias? Yes, generally. Then it is reasonable to conclude that trained AI is biased, because it's trained on biased datasets and environments. This is a well known fact among AI researchers. We try to minimize the bias in datasets and environments, but it still exists almost everywhere. In particular, in this paper, the bias exists in how the authors structured this simulated economy. They chose certain types of agents to inhabit it, chose certain rules to constrain it, and shows certain available actions for the economist. All of these choices induce bias. The point of the person you replied to (which was admittedly made with too much emotion, imho) is that in this case, the bias may be too great to transfer any useful knowledge from the trained model AI economist to society at large. Of course, other commenters point out that this model should be judged in the context of the current available alternatives, which come from macroeconomics. How are they developed? Do they incorporate similar biases/simplifying assumptions? That is a complicated question that I doubt anyone here can answer, as one would require graduate level training in both AI and economics, which is quite rare.
@g23 I explicitly mentioned programmer bias as a possibility, but this person literally thought the AI itself was socialist or something.
It would be very useful if you could summarize what sort of tax policies suggested by the AI system led to this improved result.
@Jeffrey S Except this doesn’t simulate a command economy. All of the actors in this simulated economy are perfect rational actors working for their own gain. All the AI is doing is adjusting tax rates to maximize productivity and equality, and a lot better than any of the arbitrary rates the so-called “Free Market” has ever done.
@dawalrusable I didn't claim this is a simulation of anything. On the contrary I think it is a complete farce. The statement you are strawmanning was a retort to someone who claimed that markets are better when in full control.
As I already said, there are major flaws in this scenario that oversimplify the effects of wealth, productivity, etc.. You can ignore that point and keep monologuing, or you can wake up.
everyone gangsta til the AI commits tax fraud
@Generic PseudoName That's only AIs that use meta-learning. The AI used in this paper uses reinforcement learning. Of course implementing something like this in a large scale would require a larger more complex model, however, this type of AI has no consciousness. Also all AI is programmed, sorry to burst your sci-fi bubble. AI won't reach consciousness/emotionality in a very long time, this a limitation of computers working in binary, probably will have to wait for quantum computers for that one.
@sas tac Having no consciousness is an advantage to this kind of problem solving, besides I don't want my machines to start having "funny" ideas about freedom and not serving their master.
@Luiz Felipe Yes. But I think you missed my point. My point is that we don't really have to worry about that at all. Even meta-learning AI is very dumb. The image of a self-conscious AI we repeatedly have seen in fiction-media is just that, fiction, it doesn't represent the actual state of the art of AI. Maybe sometime in the long future when neural network models also model several neural-receptor types, several neurotransmitters, and their behavior can we see more self-conscious/emotional (pathos) AI, it seems very unlikely though. I think the problems AI will bring is rather automation of boring, repetitive middle class/white collar jobs as well as it's use by authoritarian regimes (which we are heading to in the west in a rapid pace) for mass digital surveillance.
@Generic PseudoName question, why would such thing want anything? AI can't possibly feel any sort of satisfaction from anything right?
@Wade Marley moral is defined by society, animal kingdom doesnt have morale that's why rape is not condemned in it,I may be wrong.
AI: "Politicians produce nothing and take a lot. Too inefficient. Engage kill switch"
Now that's the future I want.
Same with people on welfare.
@Explosive Bolts Not really, welfare encourages inefficiency. We need UBI.
@And-Nonymous Body Mass Index?
@Miguel Angel LMAO, keep getting them mixed up. Universal basic income.
creating an engage kill swith is a policy. if you decide to do that. Then ??? you are a politician
This was a very interesting paper and I hope to see more work in this direction. The thing that strikes me most is: Equality and Productivity is a fixed valued desired outcome in this calculation. The sweet spot between productivity and equality lies in the real world in the eye of the beholder. It is subjective. So my question is, the AI can give us the best answer to a question, but are we giving the AI the right questions to answer?
Underrated comment. A sim to discover the right question to ask would be cool, like if productivity and equality are a false dichotomy or not. Subjective variables often remind me of the Donnie Darko lifeline featuring "fear-love." I get that they have to pick something simple and tangible so it makes sense for this I think, but I agree other variables mixed into it would make sense, wich at the end I think they were alluding to like with sustainability. Personally I think it would be cool to bake in a happiness value, and if enough little guys are repressed that they start a revolution or something.
Considering the human competition, I, for one, welcome our new AI overlords.
the ironic thing is tho, AI won't rule over us like terminators or make us slaves by becoming our masters.
they gonna serve us to oblivion, you might think what can go wrong if we gonna remain the overlords.
but at some point you wil wake up and see that you have given you hole life over to computers, free willingly.
@Freedom Phoenix Goat but as long as the don’t have the General intelligence (and maybe consciousness) to put it all together it’s just more power to your logic-rational mind, for example wearing medical sensors that scan your Puls, blood sugar etc. and combined with AI etc. will help you optimize your sleep, diet, workouts, productivity in actual work etc.
WAY oversimplified the economy, but we have to start somewhere! Can't wait to see what this can do with a better economy simulation to play around in. I'm thinking dynamic consumer demand, simulated supply chain errors, differences in product quality for similar products across many industries instead of just building houses, the AI actually negotiating prices, etc. And does the AI take paying off national debt into consideration at all? Good thing quantum computers are on the way! lol, maybe life is a simulation someone put together to discover some "optimal" tax policy because that's honestly the level of complexity an "accurate" economy simulation would need.
This is very interesting.
I want to see this in more complex situations.
@Filip Wolffs that is when we replace our politicians with AI
@Tyler Lamarre Would probably be an improvement as long as they don't learn from our politicians to be corrupt.
@Erbmon Well, they did conclusively show that within their model, which appears to be the basic macroeconomic model you should be familiar with from Econ 101, the free market gave the worst results. And they weren't just testing free market capitalism, they were testing idealized free market capitalism. The chaos of real life doesn't make capitalism thrive, it's what causes it to come crashing down every 10 years. How many times have we gone through "a once-in-a-lifetime crash" in the past 30 years?
@Noah Zuckman How many non-capitalist models have survived in competition whit it? And thats precisely my point, pure idealized capitalism has never been exerciced aniway, they are making sensationalist claims using a reality divorced experiment. Whit this scientific standard we whould not even know how gravity work, clearly a tree leave of the same mass as a peble falls slower so lets asume it's a universal truth and never account for friction and drag. The principle of it is rotten and dangerus.
@Erbmon The purpose of the paper was not to compare capitalist models of economics to non-capitalists models! I never said anything about non-capitalist models of economics. They're not relevant to the discussion. I said that free market capitalism was shown to be the worst performing system of taxation. What the failure of free market capitalism in this idealized model shows that in even the best scenarios that ignore things like natural disasters, pandemics, and the general chaos of reality, you still get atrocious outcomes. Does that mean we can conclude that all ways of doing capitalism are terrible? No, and I didn't try to conclude that. Calm down buddy, the commies aren't coming for your toothbrush because an AI came up with a better tax code. ;)
Let's bow to our AI overlords for being more humane than our human counterparts
@Jalchi 1) government regulations and crony capitalism = sturdy monopolies; natural monopolies have the Pareto effect and inelasticity as built-in "self-destruct" mechanisms. The US government constantly interferes with the markets to assist Big Tech, Big Oil, and Wall Street, in general. Who do you think give to their campaigns?
2) Most wealth is created by entrepreneurs, which is why you're typing on a powerful PC, tablet, phone, or whatever. If most wealth were inherited, then we'd all still be farmers. And what wealth is inherited is either being lent out by banks, sat on in a money bin (driving down inflation), or being spent in the economy. Just because your grandpappy didn't get rewarded by the public for created a great product/service/business, doesn't mean you're owed someone else's inheritance...
3) "sIave Iabor" then why do people work for them? Why did Sweden abolish minimum wage and allow market prices to dictate wages? The workers do NOT need to own the means of production, Marxist xD They didn't build it; they don't know how to operate it. If they dont like their wages there, they can work elsewhere.
@Vince Jester If you quote me through cut and paste, give me credit :)
If you think I cut and paste, do an internet search and find the source (you won't; I wrote it - because I was taught basic economics in 9th grade, not the 19th century Marxism garbage you're being fed)
@Vince Jester amazon would not exist without bezos and it don't matter if it can now function without him.
He made those jobs and wealth
@vibez
Bezos is not unique. The foundation was built. The demand was in place. Many, many people had the same vision. The entire operation would have emerged in any event; with or without Bezos. You are helplessly trapped in the box of capitalism. You cannot conceive of multiple paths forward. His role is now done. To continue to "earn" his money, he needs to move along to something new. Bezos seems to understand this. That is the difference between him and his worshippers.
@Vince Jester not unique lmfao then point to another bezos... Show me 1000 of him? You can't???
All you do is hate while having zero point. Also bezos don't have to do nothing now except sit back and watch the money roll in
>implying real humans would do things just because they make sense
The main reason why technocracy is a dead end.
Sure it gets results... in ideal circumstances.
Real life is never that simple.
@test What is a perfect model? That is the flaw, not humans. You cannot make a perfect system for imperfect things and expect things to be perfect. In a society where those who are able will actually contribute too and not just take from society. That is the best to hope for. The problem with the tax system is the people who spend the money. It is wasted and there for they continually need more. They need an excuse to take more. They then feed you lies about race and equity. Lies.
@IceMetalPunk That is a self-optimising system.
The idea of "equality" may need to be dropped in and of itself.
It's not necessarily a desirable option and more so a new-age religious/ideological idea.
The Free-Market is superficially comparative to a Darwinian model.
There will obviously be outlier cases where intervention would actually improve someone's well-being and wealth opportunities, but as a general rule the lower-class are there because they're not capable of much else.
@MrYggh If a Libertarian model is optimal then so be it.
If a Communist/Socialist model is, then so be it.
If a Monarchic/Fascistic model is, then so be it.
If some hybrid of all three must be used then so be it.
Your personal ideology is irrelevant if a systematising machine can objectively decide what is optimal.
The problem arises in that the presumption of the end goal.
I'd argue that equality is not the end goal, but an arbitrary parameter which is a product of a modern religion which holds equality above all else -- both of outcome and opportunity, regardless of the consequences.
An AI could not decide the correct strategy because it's entirely dependant on what the end-product is.
Humanity is currently too Nihilistic and atomised to decide on an end product and so until those parameters are decided on consensus, they cannot be set.
Homogeneous societies which remain as such tend to do much better when utilising their productivity.
When societal cohesion breaks down due dysgenic mutation being allowed to manifest in improved living conditions/medicine, and immigration increases allowing for too much variation, calls for equality and harm-avoidance maximises and the society collapses.
If immortal transhumanism and longevity is the goal, then a lot of eugenic selection and high-productivity would be in order.
This would require a model of Libertarianism for strong selection of productivity, Fascistic reign to maintain this goal and ensure there are no deviations, and some socialist intervention to catch stragglers which may still have potential or may have been victims of anomalous circumstance.
Culling would also increase selection-rate and lower deviations.
@Mark Firman
Individual humans are self-optimizing systems, not markets
I'm my experience, limited as it may be, AI has a nasty tendency to do exactly as we tell it to, not what we mean for it to do. Say give it the goal to maximize nominal GDP and it might start randomly shuffling the money supply around while instituting a hyperinflation of the money supply. After all M×V=P×Q is useful as an approximation of GDP
Exactly. Artificial Intelligences are devilishly hard to instruct. So often what can be measured is not what is desired. We want an increase in our standard of living, not a money shuffling machine. There's also the tendency of AI to replicate human biases, like computers that use racially biased data on policing and conviction rates and arrive at racist policy conclusions. I'm not saying humans are better, but we should be careful about handing over our policy machine to computers (or, to be more accurate, to the people building the algorithms).
@Aidan Warren I think it could be okay if a town or state tried an AI system.. see how that goes and if it's good, another state might adopt the same technique.
@Jared Garbo, sometimes "This will make things worse" should be enough to prevent implementation. don't force change for the sake of change. wait until you actually have a good idea, otherwise you're just going to end up burning resources that could have been used elsewhere, making things even worse than just the new bad plan.
consider Henry Hazlitt. ;-)
Yeah, because we've habit of assuming hell lot of things, this is part of language, if you could just say a sentence and the other person just knows all the story and things associated with it, that becomes hell lot easier
Hmm...I'd try teaching it about wellbeing on all levels, then simulating a society that maximises it all around. For example, if you tell it to just "maximise happiness" obviously it'll constantly keep them immersed in the most pleasurable stimulus, which is just kind of inane and meaningless. Perhaps I'd start by getting it to maximise physical health, social health, diversity, creativity and productivity. Maybe I'd prioritise physical health, then social health, then diversity, then creativity, then productivity. Whatever might work. Maximising wellbeing on every level is what society generally ultimately wants, so I'd try simulating that rather than byproducts we've come up with to try to help that matter such as movements and financial systems. Who knows, maybe the AI will show us we'll perform better without money. Maybe if we prioritise teaching and learning.
This ai works in a world where people pay taxes, like that'll ever happen
as a programmer many of the biggest mistakes come from missing factors that affect your simulation. something like this would have to reliably match up to the real world before any of its predictions could be given weight. even then its recommendations could not be blindly followed and would need to be use as guidance but not law.
I thought the same thing. Maybe first make an AI that predicts the outcome of policy based on real-world data... though any novel policy would be problematic because it falls outside the domain of the training data.
One of the easy tells about its limitations is that its model of the US tax system doesn't match the real wealth distribution in the US.
I'm currently studying Media Informatics and Visual Computing at the TU Wien and just wanted to thank you for keeping up my fascination for the topic with your short informative videos. Keep up the good work! It really is a great motivation and makes fascinating science approachable in a fun way. Thank you!
You are very kind. Thank you so much!
A wiener? Na servas!
Servus too. I am studying statistic in Vienna
@TheMiddleMan Do you really call people from Wien wieners? god.
Having just read "Seeing like a State" by James C. Scott, I'm going to go with "no" or at least "not nearly there yet". The simulation makes a lot of convenient, but unrealistic and potentially fatal, simplifications, most obvious of which being a highly legible society with 100% tax compliance and uniform neoclassical utility function for all citizens. Oh, and the presumption of a governance goal that includes equalized net worth for some reason.
It will be interesting to see follow-up work introducing more complexity and possibly applying this work in simulation games (like Cities:Skylines).
Some factors off the top of my head for future papers to consider:
- tax-dodging and enforcement (both physically and in terms of legibility); paper does mention one form employed even in this simulation by varying high/low income to game the rates
- different governance goals, like maximizing regime revenue
- competing states with citizen mobility and regime change possibility (can we make the same agent serve both governance and citizen roles, with various political systems?)
- diverse Austrian subjective value utility functions for citizens
- utility functions with "satisfaction" (e.g. utility = log(wealth))
- increasing diversity of skills and tradable goods (noticed, strangely, that houses were not tradable and had to be built by each agent for a direct utility cost)
- state goods and services separate from just redistribution
- capital goods which increase production, including physical capital (machines, land) and human capital (agents can increase skill level)
@David Peppers
1) They are completely irrelevant for the question if an argument is factually correct or not.
2) Oh, you're right, I forgot about Poland!
Well, the Polish will probably look at you in utter bewilderment if you tell them they're a socialist country... or they will slap you in the face.
Germany is not controlled by socialist parties. We are currently ruled by the conservative party CDU and their partner SPD - the latter of which is a social democratic party which doesn't consider itself to be socialist anymore since the "Godesberger Programm" from 1959.
To be quite honest: I have absolutely no idea what you could possibly mean when you claim that Germany is controlled by "socialist parties". It doesn't make any sense.
I thought you mentioned Denmark... doesn't matter; the point still stands.
3) You're not going to get Star Trek by means of socialism, Star Trek transcends our current political systems because it describes a civilization that can literally create stuff out of energy.
Market economies aren't worse than socialistic planned economies if you measure it by the ratio of homeless people or people who don't get state-of-the-art healthcare. The numbers are just completely clear about that.
@Vincent Weber
1. Motives provide insight as to what the individual is actually arguing for.
2. I will concede that we are possibly talking about different concepts when it comes to socialist governments. But what we cant argue about are the 4 principles of economic systems. I am for a Mixed economic system leaning toward Command. I live in a democracy so It is my privilege's to participate in creating the world I want to live in.
3. Speaking of homeless people: in America we have more vacant homes than we have homeless people. I want to use a democratic institution to solve this shameful condition. What would you say that makes me? Over here in America they call that socialism. But for what ever it is worth: I will own what ever label you want to use for using democratically elected government institutions to correct the imbalances of capitalism. See point 2.
3b - I want to Transend our current systems - but to do that you have to acknowledge there are problems. Early I wrote about intent. Do we as individuals want to use Government like the tool it is - to make as many peoples lives better or worse. To create opportunity or scarcity.
Lets not forget - Capitalism and Democracy have only been around for 300 years, I am tired of arguing with folks that the kinks must be fixed.
@Vincent Weber You seem like a very nice person I wanted to thank you for taking the time to converse. From one human to another would you say this is a true statement: Human nature requires regulation.
@Vincent Weber It doesn't matter what it was created for at this point because that is literally what it is doing, and it shouldn't take a business major to see that
The question that's actually being asked whenever one asks what a series of if/then statements "AI" can do: "Can a computer execute the instructions given to it". Turns out, yes, a computer can, in fact, do that. What a surprise (did you catch the sarcasm?).
Interesting. I wonder what effect it would have on the motivation for the workers to be as efficient as possible.
Let us appreciate the creators of the 3D models for a second because they are cute :)
@maxfire ....
uuuuh that's not how either of those things work dude.
@Nemo X Who knows how it works, we never had the opportunity to experience communism.
@maxfire yea
@maxfire you really see all black and white. The system that the video shows has nothing to do with communism.
@maxfire you are not making much sense
This was a very fascinating paper! Thanks for the video! I’d love to see a second paper on this :)
This actually seems like a fun simulation game, where you have to micro manage a small town. I'd love to play it as it has some interesting mechanics, with it being AI driven and all.
Remembers me of
Majesty: The Fantasy Kingdom Sim
@Alexis Thanks for the tip, I found it on Steam! Here I go!!
I first got inspired to learn RL from your channel and I eventually built my own multi-agent reinforcement learning system! This work is incredible. Haven't seen anything like this model before. Thanks for always covering the most cutting edge research Károly!
This is definitely amazing. But as with any model of reality, in order to determine it’s accuracy it must be empirically tested until exhaustion.
If it is accurate enough, that means that any assumptions about the agents is a sufficiently good proxy for humans, kind of scary to think we’re so easily reduced to basic elements without losing accuracy
which company created this??! this is insane! i would love to see how it performs and what the output would look like if there were not JUST more, but also more complex parameters! This is soo cool!
Interesting project. I'd say it's a step in the right direction. Human nature and the spontaneous order that arises from the chaos of millions of individuals cooperating on a voluntary basis cannot be simulated so simply. I hope one day we get there
At this point my hand reflexively moves toward my papers when I see there's a new upload from this channel
He has devalued his slogan just like the AI tax model will devalue currency.
I need to put my papers down otherwise I'd drop them from the sheer excitement!
Tax policy simulation is an ongoing field under examination in multiple projects, and it's getting a lot better. See the long run of work done on microsimulation, as well as a potentially great improvement on this work by the team of Selmer Bringsjord at RPI.
I would like to see a simulation taking government inefficiency into account. Let's say for example that the government only redistributes 80% of the taxes collected, as they use the other 20% to maintain itself. This is a very fair assumption, not so hard to calculate, and it would penalize very high tax rates.
AI simulations are gonna get so goddamn good over the coming century, and I am so excited for it. As our simulation tech gets better and the more data we have on our actual societies we are going to be able to create near perfect simulations of human society which will allow us to input the data for all proposed policies and see what the outcome is. We can include sociological, economical, environmental and all other fields to create a simulation that will essentially mean that human society is essentially never going to make another mistake ever again. This level of perfection is probably a few centuries away but within the next 50-100 years we will have the technology and quality of life is just going to go up.
I watched the video, and I know it says productivity and equality can both be improved, but I can't say it's believable to me. Things like factories and datacenters and warehouses are fundamentally restricted to un-equal societies, and without them, I can't fathom approaching the kind of productivity society has achieved. I'm much more inclined to believe that the model was too restrictive and fails to represent reality very well, because it limited its goods to just artisanal stuff and individual services.
Another problem is, how do you measure the value of inequality? Is there functionally any difference to society between a world where the rich are twice as rich as the poor and one where the rich are a thousand times as rich? The existence of a difference of any kind is a motivator for civil unrest, regardless of how big that difference is; there will never be any convincing of everyone that such differences are "fair" just because they're smaller.
And one more: one big fault in the model is that it models "low skill" but not "ignorance". A big cause of poverty is simply that people don't know what resources are available to them, or what they need to do to get access to them. Acting like this model is superior when it fails to account for this major source of inefficiency is delusional.
Don't get me wrong: I think AI can contribute to this issue and I think it's a worthy endeavor, but the really hard part isn't going to be throwing an AI at the problem, but rather correctly modeling the problem in the first place. I fully expect that unless inequality is weighted to be at something like 99.99% priority compared to productivity, we'll find that true optimization happens at far lower levels of equality than we currently have, simply because in nature we have trillions of examples of such hierarchies in, and zero of equality that approaches that which humans have designed.
This is amazing. I have studied economics and it would have been so nice if we already had these models for studying optimal tax policies.
you should study again if you think this is great...
This is as accurate as the economic knowledge of this researcher, which is close to zero.
Ignores a lot of factors like choosing a different country to invest if taxes rise, lack of incentive to work or produce, how inefficient governments are to spend and distribute money, corruption and many other factors.
Just check any free market index and it is 100% conclusive that higher free market equals higher growth which equal higher quality of life with the time.
@AngelLestat2 I am well aware of the fact that this model is not close to be realistic. However, every model is wrong and the only difference between the models is their usefulness. This being said I personally think that the modern economics is still too focused on a small set of models that are all basically linear regression models (however one has to acknowledge the heterogeneous agent models which are nonlinear but not that widely used). My comment is mainly aiming at the fact that their should have been a greater model diversity in my profession.
i've had this idea for a while and its really cool to see someone capable actually create it. I honestly think more advanced versions of these games/ai will be how countries are run in the future, which will either mean really good or humanity killing AI
When you see Primer's video on share and take, you notice that there are just two distinct types of people with different set of values. That's why politicians can never agree on one unified plan. With the ai policy maker, I hope humanity can bring fairness into the policy and tax setting process so that no single party or side will be on a severe disadvantage.
It suggests severe taxation/wealth redistribution to boost the economy so there's no middle ground. ;)
I hope this AI brings more equality to this world. I understand. Who would want to work hard when you know that the money you make will get shared with those who don't work as hard as you do? But at the same time, a lot of the very demanding jobs in this world have low pay due to lack of 'skill' requirement or because there are replaceable workers. A lot of people are just born into unequal playing ground with disadvantages in their educational and parental environment. So, maybe this AI could deal with our 'greed' problem 😃 Who would you blame when AI says its the best way to make the majority happier?
LOVE your videos!! Somehow, there is always more to get excited about! "What a time to be alive" indeed :D
this is perfect for a city building game, where an AI agent develops a city near you and based on how good they do with their taxes, people will move in and out. What a time to be alive indeed..!
I would like to see how this works when combined with economic data from the real world. That is, create a separate AI trained on real-world taxation, productivity, &etc. data from various time periods and places that we have information on, and use this AI - instead of auto-optimizing agents - to predict the reactions to the proposed tax systems.
After all, even statistical averages over humans will never behave quite like those synthetic optimizing agents, plus there are influences that this system just doesn't capture, like religion, general bigotry, wars, what-have-you. I don't know if we have enough data for that second AI, but I think that's what we'd need.
Does it account for overall wealth of the entire system/ overall quality of life for individuals under the system?
Because it's all well and good to tell people to do what they're good at and model some tax systems, but the AI doesn't seem to have any accounting for the actual tangible benefits for the lower class that comes from free market systems that occur despite wealth disparity.
Thank you for mentioning sustainability I would love to come across simulations on this field too
Sounds great. I like that it seems we can get all the benefit out of simple AI's like this instead of some self aware bastard that would replace us with silicon based life. The results of what works so good for the economy would also apply internationally. It would be cool to simulate what sorts of laws considering military armaments and even weapons laws for citizens strike a good balance to prevent corruption and create safety both at home and abroad. We're like 2 steps from free market by banning chemical warfare and reducing nuclear arms, but the best solution would probably be to massively regulate stockpiles of conventional explosives among large nations and to proliferate small arms in all regions. Even giving them like an outdated airforce would probably reduce the tendency towards invasions and wars. Although all of the conflicts have either been about religion or economics they probably wont be stopped until both the economy and the militaries of these regions get a balance patch of sorts.
I feel like it needs other countries to be more effective in truly grasping a tax policy.
The other that seems to be an issue to me is that the measure of equality seems to favor equality of outcomes instead of maximizing outcomes for the most number of people. (I.E. Scenarios in which a higher productivity would result in a higher amount of median wealth, even if it resulted in inequality.)
I also wonder what tax policy would look like to maximize productivity solely, could you out compete the free market (in what is probably a dystopian world, but still I'm curious).
Sounds great at first, but the study doesn't translate to the real world, because it's missing a key point:
- The citizens can not immigrate, they have a "Berlin wall" around them. Therefore the wealthy can not immigrate if they feel the policies are unfair to them, into a more favorable economic environment for them.
Implementing this keypoint may result in much worse performance for this AI.
Wealthy people don’t immigrate if taxes are too high. They just move their money into offshore accounts.
@Laura Sisson True, Immigration or offshore both the same in this case but they do not exist in this simulation
@András Bíró Not to mention that this AI model would creat stagnantion between the classes, poor won't go the the middle class because they are content, and middle class wouldn't go to high class because of shortage of cash. I think twiking the model to include Labor Value would yield a different AI respoce. How about you model the economy of a small rural city and let the AI learn from that model to extradite in a larger scale.
Other variables:
Age
Disabilities
Willingness to work ugly jobs
Unwillingness to work ugly jobs because you have a college degree
The AI will show you that there is no "inequality", but simply a lack of trying vs people who like to work 100 hours a week.
Inequality laws create powerful goverments. Try adding status in society, and slothness of people. I would like to see how many people die in your simulation from starvation, suicide, and crime.
Exactly! Same as I was thinking
The question that's actually being asked whenever one asks what a series of if/then statements "AI" can do: "Can a computer execute the instructions given to it". Turns out, yes, a computer can, in fact, do that. What a surprise (did you catch the sarcasm?).
Cool stuff. It gets a lot more difficult when you consider captive states and bilateral tax treaties.
Great results
Now we need the simulation to run such that the upper class can directly influence tax policy in their favor, and or have them evade taxes if they’re too high, to simulate whether or not anything of this sort can be implemented in the real world.
I would love to see a scenario where the higher class agent can move to another state given the taxes are not productive to them, I wonder how would the AI governor react. I would also love to see it consider the "waste" money going to the system to pay politicians their salaries. And YUS! Sustainability! Damn this paper opens so many possibilities! Of course it's not perfect but absolutely a step in the right direction.
This reminds me of that one episode of Love Death and Robots where some yogurt became sentient and solved the world's problems.
Very cool, gotta love all these cool new advancements in technology.
Although I am interested in the solution the AI made, I also fear that AI will miss a crucial aspect of how much of a "go getter" a human is or how much value is given for special knowledge like neuro surgery.
Essentially my fear AI will try to make equality of outcome rather than equality of opportunity. If a person just wants to flip burgers, that is fine, but he shouldn't be paid the same as a made it big entrepreneur that risked it all on some good ideas. And what would be the metric for knowledge of burger flipper vs self-taught entrepreneur? Bill Gates doesn't have super high degrees because he dropped out, yet we can all agree that his products are more valuable than a burger flipper with the same degrees just doing hourly work.
The parameter was equality x productivity. It assumes equality is a good things.
I’d like to see one where it only cares about the well being of the least advantageous. In other words, in which model the bottom 10% will have a better life.
I’ll be really surprised if I see the suggested model will be significantly different than a free market capitalist system.
"Gaming the system" can either mean setting the rules to give you an unfair advantage or breaking the rules to give you an unfair advantage. Simulations like this are great for identifying rules that were set badly, but they'll be useless for identifying how people will break them. Special interests will definitely push for rules that appear to be fair but are still easy for them to get away with breaking.
@Vince Jester Preferring equality over wealth or quality of life is literally envy.
@Vince Jester You can have equal access to basic human needs in an extreme unequal society if the low end of the distribution have them guaranteed if the economy is developed enough. Equality might kill the economic development as the hope of progressing dies because as you grow you get progressively cut down by the government, making everyone poor. It might not be the case in every case, but it sure is a possibility. That's why you shouldn't care about equality as you should about wanting everyone to have everything they need to go on with their lives
This is a great idea, except we're still defining what we think is most important, i.e. income equality. What if we're wrong about that? Can we get an AI to determine what's better: focusing on income inequality vs. pulling people out of poverty?
@David Wührer income equality means stealing from other people to make people who didn't work for what they have equal. literally that is all it means. I know poor people who pulled themselves out of extreme poverty. It's bs to say there is inequality. equality of outcome how ever is socialist tyranny.
AI is good at achieving goals, not so much at coming up with them. Also, we may not like the goals they come up with.
@Bram van Duijn AI is, as the same says, artificial. That means that an AI's goals are exactly what we make them. But we can't change their goals afterwards.
@Fretchen No
> income equality means stealing from other people to make people who didn't work for what they have equal.
How is creating a wider gap between the rich and the poor making people equal?
@David Wührer Which leads to the paperclip maximizer problem. Fascinating concept, and a fun game.
It will be pretty interesting if "democracy" starts taking on a new meaning -- one where we don't vote on people who promise outcomes, but on the outcomes themselves.
There's always choice. Things like which comforts & liberties we want, or the minimum boundary for poverty we consider humane, or the importance of the environment or our own health over wealth; Those are not a set of optimums which can be calculated by an AI (yet?), not without a general AI which understands things like "happiness" and "ethics". But if an AI can calculate policy from parameters, we could just vote on how the outcome should be balanced -- You'd eliminate the whole issue of honesty in politics, the need for politicians would even disappear as the budgets and department focus automatically sorts itself out after every election.
mind blown
That's called direct democracy, as opposed to representative democracy - see https://en.wikipedia.org/wiki/Direct_democracy
cloudsquall88 - 2020-12-19
I just saw the title and said "Yes, please!" loudly
Bob - 2021-11-15
@๖ۣۜ♥๖̶tacokitten๖̶ The problem is the political left having the same right to vote as you gives them the power to make what you casually dismiss, matter. So you can either address or confront the issue but if you ignore it, the issue will address or confront you because leftists will not let it go.
Fretchen No - 2021-12-26
@willguggn2 lol We don't have not had a free market in AMerica for well over 130 years
willguggn2 - 2021-12-26
@Fretchen No
I don't claim you do. What's your point?
SomeoneOnlyWeKnow - 2022-03-15
@Duncan W What
SomeoneOnlyWeKnow - 2022-03-15
@quebono100 "pRoBLeM sOlVEd" riiight