Centrally Planned Economies Didn’t Work – Algorithmic Planning Could Change That
Central planning was tried in agrarian economies but didn't work. What about economies that have already industrialized? Algorithmic planning could give central planners a new lease on life.
Right now we think of artificial intelligence as ChatGPT, Claude, and Grok. Kind of like Jeeves – but smarter. AI is just a chat-based interface we use to access vast amounts of information stored on the World Wide Web.
While this is the primary context in which the average person uses AI now, this isn’t how we’ll always use AI. For many leading technologists, the goal is to make AI omnipresent and autonomous. Eventually it’ll drive cars, replace workers, find its way into humanoid robots, and eventually govern us.
The field of artificial intelligence isn’t just about the science of how computers work. There’s a philosophical element to it too. AI is also about developing inorganic cognitive reasoning. To do that, you need to understand how humans make decisions. For many technologists, the goal isn’t just to make AI sentient, it’s to replicate human intelligence so that AI can liberate us from our own fallibility.
AI planning is a subset of artificial intelligence that trains AI to make its own decisions. It relies on developing a sequence of actions – a plan – that AI can follow to achieve a pre-planned goal. Within the parameters of those actions, AI can adapt, developing the capacity to think while teaching itself how to make better decisions.
Let me use running as an example to illustrate this. I’m a runner. My goal is to run the Boston Marathon. There are parameters I follow as I work toward accomplishing this goal: I adhere to a training schedule, manage my diet, and run qualifying races. As I track my performance within those parameters I observe whether or not I’m making progress toward my goal.
AI planning works in much the same way. It allows AI to expand to multidimensional interfaces and adapt to real world conditions through smart algorithms. As it adapts, it learns from its actions, gaining more and more command over its decision-making faculties.
AI-powered smart algorithms will fundamentally transform how humans make decisions going forward. Humans won’t have to expend brain power, time, or financial resources to accomplish goals. And they won’t be burdened by the limitations of human emotions and ego either.
Every institution where human decision makers are at the helm will be able to use smart algorithms and deploy AI to achieve specific goals. If a business wants to hit a specific earnings target or the government wants to reduce unemployment, it will have a more efficient tool to do so.
While there are a number of benefits to this, it raises questions, especially for governments. Government exist in large part, to provide collective benefits to society that we can’t provide for ourselves individually. The government extracts resources from the population – often through personal income taxes – and provides benefits such as infrastructure, defense, and these days a whole host of welfare programs.
But what role will the government play in an automated world where algorithms and AI – rather than humans – are responsible for making decisions? If systems are implemented where human decision making is removed from the process of governance, what’s the point of electing representatives to advocate for us in the first place? Can democracy and AI peacefully coexist?
More importantly, when applied at scale, how will algorithmic planning change the economy? When the market is planned and humans are assigned a role to play in it, how will humans maintain their autonomy respective to the artificial intelligence appointed to manage them?
Fortunately, we don’t have to look too far back in history to see how this could play out. That’s what this essay looks into. It will dive into historical precedent for how central planning was used to shape outcomes within an economy. It will highlight the shortfalls of planning to foreshadow some of the limitations AI-powered algorithmic planning can expect to face in the future.
It will argue that if algorithmic planning is to be successful, human economic participants will have to surrender immense amounts of personal data, often against their will. Instead of existing in a market where humans have some degree of choice, the AI-powered economy of the future may increasingly rely on surveillance in order to manage the outcomes AI is programmed to achieve. By understanding the trajectory of autonomous decision making we can take action today to mitigate the unintended consequences of AI proliferation tomorrow.
This essay dives into:
⚡Why AI could displace humans in the decision making process
⚡Why central planning failed in agrarian economies
⚡How the U.S. economy already practices a degree of centralized management
⚡How algorithmic planning can make central central planning easier for modern economies – and the cost they’ll pay
☕ Thank you Tomorrow Today subscribers.
Your support makes it possible to share thoughtful commentary like this about how the world is rapidly changing and the things you can do to prepare for all the changes that lie ahead.
Centrally planned economies replaced individual decision makers with government experts in much the same way AI will likely replace human decision making in the future. Historically, these economies failed to properly allocate resources, resulting in mass starvation events.
A centrally planned economy – also known as a command economy – is an economic system where the government – rather than the market – makes decisions. The government owns the means of production while bureaucrats working within the government set prices and determine what should be produced and in what quantities.