Automating Government, Global Economics and Democracy

I recently completed a course in Machine Learning from coursera.org offered by Stanford University with Andrew Ng as the professor. The course covers many aspects of machine learning, its purposes, including pattern recognition, data compression, and others,  various mathematical models including linear and logistic regression and neural networks, and the actual programming required for modeling, testing and validation. It has been a long time since I programmed anything very substantial so I was very happy to learn that the higher level languages that are now commonly in use handle matrix math like a breeze allowing for the development of very sophisticated neural networks with relatively few lines of code. Not very long before that, I took another course from coursera.org on Global Macro Economics. The diversity of the courses that I study is not simply due to curiosity.

After having observed how people handle the economies of the world and watching the results in terms of happiness, economic well-being, distribution of wealth and many other factors, it is clear that people are doing a mediocre job at best of governing. When you think about it, the reasons for this are not very surprising. When we elect representatives to our provincial governments, the Canadian Parliament, or in the US to the US to Congress, the Senate or the Presidency, we hope that these representatives have some idea of how to run a country. The reality is that most are from various walks of life and many of the higher level representatives have law backgrounds. They know very little of economic, environmental or psychological studies on what impacts the relative well being and most could certainly not be considered experts. No doubt their political life is an education unto itself, but what they learn is not necessarily based on any reality that impacts the masses. More likely their education is based on political influences.  So why on earth would we expect a very positive outcome of a mediocre democratic system?  Even the diversity of viewpoints which could be theoretically expected from a diverse group of people is generally stifled by party leadership.

As we are moving towards a world where artificial intelligence (AI) is utilized in everything from accurate medical diagnostics to ever safer and eventually self driving vehicles, where sensors are becoming ubiquitous in society and measures of everything are being taken, doesn’t it make sense to automate some of the functions of government? While I am not suggesting that we do away with any government, if we were to simply set the parameters of where we want society to go, and the many variables  that shift the course of economics and society, perhaps we can also create machine learning algorithms that can steer the economy more successfully than our governments have.

If climate can be modeled with some degree of accuracy and auto-driving cars can successfully navigate the complexity of safe driving on busy roads, the analysis of “Big Data” pertaining to government functions, political decision making and economics is surely not very far out of the grasp of current computer analysis and machine learning algorithms.

As an individual who understands the potential in Machine Learning and the basics of Global Macro Economics, I recognize the potential for growth in these combined fields of endeavor. Projects like the “Brain” project and other AI initiatives will undoubtedly improve the tools, the methods and the capacities of our AI engines. As we apply these ever more powerful tools to an ever more complex world model that is becoming increasingly difficult for our political systems and representatives to manage, perhaps we can find some new and interesting answers to questions such as: how do we eliminate the extremes of poverty, increase global happiness and develop better health care and educational systems and policies. And how do we do that while ensuring sustainability with a low ecological footprint? Undoubtedly, if we had the capacity to evaluate all of our historical data from the perspective of sequence of events, we would discover new relationships between initiatives and outcomes that might help us to steer the future course more effectively.

Like every other use of AI, such an application would need to be structured in bits and pieces at first, taking on the evaluation of specific areas of governance. Eventually, however, with access to far more information and measures than any one person can possibly make sense of, AI systems would be able to manage certain government functions more effectively than our elected representatives and would be able to respond to global shifts more effectively and safely.

As a relative novice in both the fields of AI and Global Economics, I would love to hear the comments of those who have more experience and expertise in each of these fields. A collaboration of government, scientific, health and economic experts would be a welcome start to such an initiative. For those who may already be aware of existing initiatives, I would love to hear from you, so please contact me at my e-mail: garth.schmalenberg@gmail.com

Garth

 

 

Comments are closed.