There is a myth based on experiments in the 1840ties that a frog would react to rapid changes but not to a slowly warming water. The myth contented that a frog thrown in boiling water would jump out, but a frog sitting in water that slowly is heated up would not jump out and eventually get boiled.
This myth is often used to argue that humans are like frogs: they only react to rapid changes, like the ones expressed in disaster. For gradual changes, they don't react until it is too late. This comparison of human behavior and frog behavior is flawed for several reasons.
First of all, the myth is flawed in two ways: throwing the frog into boiling water will simply kill the frog - no jumping out. The frog in slowly warming water will get more and more active as the water temperature gets closer to the thermal limit of the frog. Changing location is a built-in natural thermoregulation strategy for ectotherms like the frog. It is necessary for survival. A frog that is gradually heated will jump out. The frog is programmed to react actively when thermal conditions get closer to the limit it can tolerate. If the frog has a way to jump out of the slowly heated water, it will jump out: it is programed to do that.
Second, the comparison of behavior of frog and humans implicitly assumes that both are the same type of agents. This is not the case.
Despite clear evidence that the myth of the frog waiting to be boiled is flawed, this myth is often used to argue that like the frog we would react to rapid change (like disasters), but we would not react to changes that slowly make our situation worse (like modern climate change) until it is too late.
Stating that humans are like the frog in not acting to slow changes is fundamentally flawed because frogs and humans are very different types of agents. The frog is very close to a rational agent. Humans are goal-based and utility-based learning agents. The behavior of these two types of agents are not comparable.
The frog is programmed to react actively when thermal conditions get closer to the limit it can tolerate. It is programmed to do something in extreme situations. It doesn't have a choice: if the water gets too warm, it will jump out. It is programed to do that. If a thousand frogs sit in slowly warming water and they all can jump out, then they all will jump out. That's what the programming requires.
Not so for sapiens. Some humans may not react to slow changes until it is too late. The reason is that we are programmed with the freedom of choice. We are not programmed with inherent limits and fixed decision making if we approach the limits. As a result, we often exceed our limits.
All agents can be described as having a percept of the environment through sensors, having means to make decisions on actions, and having effectors to implement the actions and impact the environment. Based on how percepts lead to actions, we can distinguish at least four main types of agents (Figure 1):
In our complex world the individuals in a community can have a wide range of goals, and often the goals are conflicting both at the individual and the community level. Likewise, there is a wide range of utility among the individuals and between different communities.
A good example of goals are the seventeen Sustainable Development Goals (SDGs). Some of these goals are conflicting with other goals, and they also are conflicting with societal goals not included. Reasoning could be used to weight the likelihood of achieving the goals against their importance. However, humans tend to bring preference into the decision making to achieve maximum utility.
Humans appear to be hybrids of goal-based and utility-based decision making. On a wide spectrum, individual humans have unique combinations of goal-based and utility-based decision making. This leads to a wide range of psychological and social behavior.
In recent times, a terminology is used to put individuals on a spectrum of neurodivergent to neuronormal. It might be better to use a spectrum from goal-based to utility-based decision making to characterize the agent type a particular individual represents. It seems like those agents that are more goal-based are often considered neurodivergent, while those making decisions based on utility are the neuronormal ones.
Humans have the ability to switch from goal-based to utility-based agents and visa versa. In the movie "The Triangle of Sadness", the super wealthy passengers on the luxury cruise ship obviously are utility-based agents with all decisions resulting from maximizing utility. The staff are more goal-based agents who do what they do because they must do it to keep the job and make a living. After the disaster that leads the ship to sink and one staff member and several passengers to be stranded on a seemingly deserted island, both agents switch type: the passengers are now goal-based agents with the only goal being survival in a situation where they have virtually no skills to facilitate this survival. The staff member has all the skills necessary to survive, and she can make choices based on her preferences to maximize utility: controlling those who were controlling her before the disaster and getting maximum satisfaction out of this controlling.
We humans come with far less rational or goal-based (survival) programming that would limit our choice of action and would guide our decision making. We enjoy the freedom of choice this gives us in our decision making, but this also gives us the responsibility to actively look for the limits and not cross them. Unfortunately, most people are not aware of the fact that freedom to make choices comes with the challenge of making the right choices. In a complex world, this is a very difficult challenge.
Concerning the planetary emergency that our modern society has created, many of us can clearly see where we should "jump" to tackle this emergency, but then don't act. Many prefer to ignore the mounting risks, even as the disasters multiply (e.g., extreme weather events resulting in more hurricanes, droughts, floods and wildfires). Like the frog thrown in boiling water, many would die in these disasters.
The freedom of choice allows sapiens to make utility-based decisions. The questions asked are "What will make me happy? How do I get maximum rewards? Sapiens often don't ask the (goal-based) questions like "What do I need to do to reduce risk?", "What would ensure maximum benefits for all?", "What do I need to do to ensure our survival?"
In modern society, there is a strong focus on individual freedom in decision making. This freedom often results in utility-based decisions. These decisions do not reflect sufficiently the necessities for survival. What makes us happy in the short run (utility) is often not what is needed for survival. As Barry Clemson pointed out to me, addicts, people who consistently overeat or live on junk food, those who refuse to exercise, etc., are exemplars of this utility based decisions at the expense of survival.
Given that Sapiens have little programing helping them to make the right choices and decisions - or even just acknowledge that they need to think about what choices to make, it appears that Sapiens are just too poorly programmed for unguided individual decision making. What made Sapiens such a powerful invasive species is reason. Sapiens are able to reason with each other and then come to goal-based decisions. This is much more than just summing up utility-based decisions of individuals, for example in majority-based decisions such as elections in democratic systems. Having said this, it becomes obvious why societies in which the utility-based decisions of individuals are at the basis of the governance system, as is the case in democratic systems, have a high probability of not supporting the goal of a just, equal, diverse, and inclusive community in which sustainability can emerge.
Many indigenous cultures recognized this, and they had very slow and very long processes of reasoning for making decisions that would determine the future of their community. The processes we have today in most modern societies are inappropriate for any democratic system to be successful in the long run.
From a systems dynamics point of view, utility-based agents with a high degree of freedom of choice can not manage the system successfully unless they have some guidance from the outside of the system or the system itself is designed with processes that facilitate constructive reasoning. Utility-based agents need assessments of system state and trend from the outside to balance what makes them happy with what is needed for survival. They need reflective reasoning on proposed decisions from the outside.
What would be a governance system that accounts for what we know today about the human brain, about our cognitive biases, about the agent type we are? Whatever the system, it ...
With the advent of AI (i.e., artificial intelligence, or, as Harari prefers, alien intelligence), we need to ask the question what type of agents the AI agents should be. If they are goal-based, who will determine the goals that will determine the decisions they make? What rules will they have to follow in their decision making? Will they have the freedom of choice that we enjoy?
Barry Clemson commented on this: “Harari says that the social media companies gave their AIs the goal of maximizing user engagement. From there the AIs discovered that blood and gore and hate and conspiracy theories maximized user engagement. Thus the humans (Zuckerberg et al.) did not explicitly decide to make social media into a cess pool ... they "merely" followed the mandate to maximize profits and the AI did the dirty work. I have elsewhere discussed why maximizing a single variable always leads to bad outcomes.”
This comment makes clear that just giving goals to AI may lead to reaching these goals, but the AI may choose pathway to get to the goals that are in violation of our normative as well as descriptive ethics. Alien intelligence does not have the limitations we have based on our ethics and may explore pathways to goals we would not even be able to imagine because of our inherent cognitive biases.
Will they be designed as agents that are utility-based? If so, what will make them happy? Will they be better able to balance the goals they have (such as to ensure a society where sustainability can emerge) with what makes them happy? Will they be more like the frog and have the rationality to "jump" when needed, or will they be more like us and in some situations more prone to "boil"? Before we engage in creating many AI agents, we might want to have a long reasoning about what agents we want to create.
sharePublished on Sep 26, 2025, 14:40:37 on blogs at P4US, COPE. Last updated on Mar 22, 2025, 11:30:47.
To submit your comment, please login first.
The purpose of a system is what it does (POSIWID). Since Stafford Beer formulated this foundational statement, a lot of effort has been on designing societal systems that do what their intended purpose is. To a far lesser extent, the statement has been used to determine the purpose of the physiol ...
Published on 2023/10/11/11/24/ by Hans-Peter Plag
György Buzsáki (2022) states that "Perception is what we do." Christof Koch (2019) describes consciousness as "the feeling of life itself." The feeling of life is the result of our perception of the inside and outside worlds combined. Then, what we do is what we feel and what we a ...
Published on 2022/10/12/09/32/00 by Hans-Peter Plag
Barry said a few days ago that the web pages of Place4Us offer too many choices. He claimed that people get confused and don't want to have so many choices.
I agree that young people today live in a world where choices are made for them by algorithms. During the time they spend with their sma ...
Published on by Hans-Peter Plag
In many publication considering the challenges or dangers of AI, the fact that AI can do many things better than humans is the basis for a reasoning that AI might take jobs away from people and increase productivity. This is a fundamentally flawed view on the impact of AI. It is flawed because it ...
Published on 2023/04/11/00/10/00 by Hans-Peter Plag
It's not about what we can do but about what we feel when we do it.
It's not about the poem we write but about what the poem we write does to us.
While Generative Artificial Intelligence tools might write poems that seem to be better than our poems, writing the poem will not create an exp ...
Published on 2023/08/27/10/24/00 by Hans-Peter Plag
Since the last more than 70,000 years after the eruption of the super volcano Toba, humans, or should I say, sapiens? have migrated. In fact, sapiens evolved into the most invasive of the large mammals, the only one found everywhere on the Earth's surface. Sapiens ...
Published on 2025/03/02/19/16/42 by Hans-Peter Plag
There is a myth based on experiments in the 1840ties that a frog would react to rapid changes but not to a slowly warming water. The myth contented that a frog thrown in boiling water would jump out, but a frog sitting in water that slowly is heated up would not jump out ...
Published on 2025/03/08/20/24/02 by Hans-Peter Plag
Rules of Conduct
Search in all VCCs: No Yes
Submit Selected