User:CognitiveMMA/sandbox/GCI4MODELS2022/modelingsocialgood

Modeling Collective Social Good

As the old adage says, you can’t achieve what you can’t measure. In order to measure any property, there must be a model for what that property is. But how can collective social good be modeled? It is clear that there are a great many methods for measuring social impact. What is less clear, is that the choice of the means by which social good is modeled must also determine which problems are not reliably solvable, and that any model for social good that can’t measure itself is one that can’t reliably identify those unsolvable problems. Luckily, Human-Centric Functional Modeling offers a potential solution.

Human-Centric Functional Modeling attempts to understand systems through defining models capable of providing complete representations of the system's functions, thereby removing any need to understand how the functions are implemented. Just as the definition of physical space accommodates all possible motion vectors without having to specify any particular one, HCFM defines “functional state spaces” to provide this complete representation of the behavior of any system within a given domain of that system’s behavior. This technique has been used to define a “human state space” as a composition of a physical state space, an emotional state space, a conceptual space for the cognitive system, and an awareness space for the consciousness. In addition to the state space defined for each system, each system also has an overall “fitness”, measured in terms of ability to execute all of its functions. This overall fitness can also be seen as the measure of the “well-being” of the system.

The existence and validity of this internal measure of physical well-being in terms of fitness to execute all of our functions can be confirmed by a simple experiment that assesses whether individuals are able to reliably assign a relative measure of well-being to different physical states. Assume that in the first physical state, the subject experiences a runny nose, perhaps a slightly elevated temperature, but can walk, run, and perform all of their other functions adequately. Assume that in the second state, the subject experiences bleeding from all orifices, a highly elevated (feverish) temperature, and is severely limited in all of their other functions. Even without naming these two physical states (the first being the common cold, the second being Ebola), most individuals reliably choose the first state as correlating with greater physical well-being.

Assuming that every system in a living organism has zero fitness when it can no longer function, living systems might be expected to vary in fitness over their lifetimes. Human-Centric Functional Modeling represents systems within living organism as moving through their functional state spaces in a way that ensures their fitness stays within a stable range. In other words, as systems move through their functional state space, they also move through a “fitness space”. If the behavior of any living system (e.g. the body, emotions, mind) can be seen as moving through its functional state space, that motion through functional state space can potentially be seen as the result of a continual process of optimization in fitness space. In order to achieve this continual optimization of fitness for all possible actions in a given functional state space, HCFM defines the most general possible “fitness space” so that it applies to any action in that space. This fitness space is defined by the current fitness of the system, the target fitness of the system, and the predicted fitness as a result of executing a prospective action. General problem-solving ability in a collective intelligence is hypothesized to be achievable through finding a set of equations that keeps the fitness of the collective cognition within a stable range (keeps fitness within a basin of attraction), and pairing that set of equations with an algorithm which selects the next collective reasoning process to execute on the basis of whether its position in fitness space matches the equations. The equations known to solve this problem of producing dynamics that are confined to an attractive basin are nonlinear and non-deterministic. The importance of separating the human organism into different domains of behavior (body, emotions, mind, conscious awareness), each with their own measure of fitness, is also clear. Because while the fitness (well-being) of the body might impact the fitness of the emotions, mind, or consciousness (or vice versa), each of these systems obeys different dynamics in their own fitness space.

Do Ethics and Morality Define Good?

If it is possible to describe well-being in terms of fitness of the system, then it is possible to describe what is “good” for the individual. If it is possible to define what is “good” for an individual, then it is possible to define what is good for a collection of individuals. It’s important to distinguish “good” in the sense of what serves the collective well-being, from “good” being defined in terms of any ethical or moral code. Any ethical or moral code can be defined as a set of reasoning processes that connect a set of concepts. Any set of reasoning can be represented in terms of a set of paths through the conceptual space (space of concepts) that serves as the functional state space of the cognitive system. If the above model is correct and the general problem-solving ability (intelligence) of the cognition is created by navigating the conceptual space in a way that continually adapts to produces dynamical stability in fitness space, then the fact that those solutions are nonlinear and nondeterministic implies that no set of reasoning processes, and therefore no ethical or moral codes, can possibly create the same fitness (well-being) for everyone. Therefore adherence to ethical or moral codes, or the codes themselves, are inherently unsuitable as common standards of well-being, particularly for very different cultures. In existential traditions such as yoga, this opinion that well-being must be experienced rather than reasoned out has been echoed for thousands of years.

Approximating Collective Fitness

How do we approximate the fitness of the collective cognition in order to reliably be able to decide to collectively execute one reasoning process as compared to a great many others? How do we approximate the fitness of a process in achieving its targeted outcomes? What is human well-being in terms of physical, emotional, and mental well-being, as well as in terms of conscious well-being? How do we approximate the fitness of a process in achieving human well-being? An entire Semantic Metrics Framework has currently being proposed to begin to address this issue. Within that framework, a Universal Impact Metrics Framework has been proposed to define specific impact metrics for specific initiatives. An example of the use of that Universal Impact Metrics Framework is the South Sudan School Uniform Value Chain, where collectively intelligent cooperation to select the value chain that is most fit in achieving the impact of job creation is predicted to increase the job creation per donor program dollar by up to seven hundred and fifty times. A "Social Impact Marketplace" software platform containing an algorithm which orchestrates a "collective reasoning process" to select the optimal social intervention by leveraging that metric is one possible approximation of a general collective intelligence platform.

Social Good and Decision-Making Capacity

What if being able to find the solution with the greatest social good means creating the capacity to review ten million prospective solutions per week requiring expertise in one hundred and twenty subjects, and we can only review ten solutions requiring expertise in one subject? What if not finding that optimal solution means a certain number of people will die from poverty, or from lack of access to health care? What is the cost in terms of social good of our lack of decision-making capacity? One key factor in increasing decision-making capacity is having an objectively defined model for social good so that a far greater number of prospective solutions can be objectively compared. Another key factor is having a universal model for social good so that all prospective solutions can be compared.

Social Good and Diversity

If there is such a thing as a solution that is optimal for a group, then what about diverse groups? What about conflicting groups? What about conflicting problems, that is, cases in which solving one problem breaks the ability to solve another? What about solving a great many problems at the same time (e.g. solving hunger which leads to over fishing or over farming)? What about problems that occur over time scales that cover several generations so that current decisions will affect those who aren't even here yet? The concept of using general collective intelligence does not mean being confined to a single solution selected by a collective intelligence. Instead it means increasing the collective decision-making capacity so that ANY solution which optimizes collective fitness (collective well-being) can be reliably chosen. This might entail a different solution for every single individual. Furthermore, while the reasoning of one group can conflict with that of another, and therefore any morals and ethics (which are based on reasoning) can potentially conflict, the well-being of groups can only conflict under the assumption of a zero sum environment (a situation in game theory in which one person's gain is equivalent to another's loss, so the net change in benefit is zero), in which the well-being of one group might mean the extinction of another. This might occur, for example, when there are two people but only enough resources (food, air, water, etc.) for one. There are mixed opinions regarding whether this zero sum assumption is correct. Some opinions suggest that this zero sum assumption is true, namely that there are global tipping points in the consumption of resources, and that combined with overpopulation this means that the well-being of some must conflict with the well-being of others. Other research suggests that while tipping points exist, those tipping points are more accurately defined in terms of the well-being of people and the planet, rather than in terms of limits to consumption, in the same way that an athlete might consume several times more food that average and still be extremely fit, while an inactive person might consume that same amount of food and be morbidly obese. Furthermore, with general collective intelligence, "green growth" might be reliably achieved in order to decrease consumption and increase available resources to the point that this zero sum assumption is demonstrably false. Outside of this zero sum assumption, modeling problems in terms of human well-being allows any number of problems to be attacked simultaneously for all people across all possible variables and time scales, limited only by computing power.

Social Good and Consciousness

There is great importance in differentiating cognition from consciousness. In Human-Centric Functional Modeling cognition is represented as operating within a "conceptual space" (the functional state space of the cognitive system), while consciousness is represented as operating within an "awareness space" (a space of awarenesses or experiences that acts as the functional state space for consciousness). The cognition is represented as navigating from one concept to another, while the consciousness is represented as navigating from one awareness to another. This awareness determines the focus of attention, or in other words, it determines which problems the cognition is directed at solving. An extremely high intelligence with a low level of awareness can easily be directed at solving the wrong problem (a problem that does not significantly impact well-being). This raises a number of questions. Can an individual's level of consciousness be increased? Can the collective consciousness be increased to become collectively more aware of better ways to define problems and to become more aware of better solutions? Can the collective consciousness be raised to become more aware of the need for greater collective consciousness where some problems must remain unsolved without it?

Moving Forward From Current Social Impact Metrics

What happens if some metric is missing from current social impact metrics? Can that problem ever be solved? What happens if each set of metrics has something missing and therefore has a different set of problems that can’t be solved, how do we create a process that reliably identifies and fixes these holes? What if we see this as a problem of morals or ethics, and our morals and ethics don’t identify and fix the problem? What are the consequences of leaving all of these challenges unaddressed?