Ethics of Process and Ethics of Outcome
In this article, we delve into the general principle that the nature of the whole and the part applies not only to matter but also to actions. From this approach, we consider the ethics of the process of action and the ethics of the outcome.
The Nature of Whole and Part
For instance, when thinking about complex systems like organisms and intelligence, we are reminded that the sum of the properties of the parts is different from the nature of the whole.
This is understandable even in simple mechanisms.
For example, the overall nature is different depending on whether two gears are placed separately or meshed together.
Even a plastic straw becomes useless when cut in half or split lengthwise.
Though they are just two gears or thin plastic sheets, their arrangement or connection can endow them with new properties as a whole.
Whole and Part in Action
The same applies to actions.
The overall nature of an action differs from the nature of its parts. The way parts of an action are combined can result in new properties emerging in the overall action.
There are two axes of decomposition in actions. One is the parallel axis. Actions performed separately by many people are parallel actions. The whole can have properties different from individual actions, depending on how they are combined.
For example, when many people travel by car, the total amount of gasoline consumed or carbon dioxide emitted doesn’t change much depending on when they travel.
On the other hand, if many people use the same road at the same time, traffic jams occur. The phenomenon of traffic jams depends on whether the parallel action of traveling by car happens simultaneously in the same place or at different times.
The second axis of action decomposition is the serial axis, the temporal sequence of one action and the next.
Take the high jump in track and field. It is a combination of running and jumping. If you run and jump separately, you can’t clear a very high hurdle.
By using the kinetic energy gained from running, you can jump higher. The height you can jump changes depending on how the actions of running and jumping are serially combined.
Ethics of Process and Ethics of Outcome
Actions produce some results. These results can create or maintain value, or they may bring harm. Additionally, results are multifaceted; they may appear valuable from one perspective and harmful from another. A single action can affect multiple things, producing separate outcomes for each.
The evaluation of an action is determined by looking at the overall picture of these results, but that’s not all. From an ethical standpoint, it’s common to focus not only on the good or bad of the results but also on the ethicality of the process of the action. Even if the result is very valuable, if the process of the action involves ethical violations, the overall evaluation of the action decreases.
This represents the ethics based on results and the ethics as a duty. Duty-based ethics judge an action as immoral if it does not adhere to the imposed duties, even if there are no ethical issues with the results of the action.
This may seem irrational at first glance. There is a rationale that if the process of an action involves an ethical violation but does not result in any negative impact, then there is no problem. Focusing only on that one action, the ethics of the process seem correct.
However, it is necessary to consider what would happen if the same ethical violation occurred in other actions of the person or others. This is the concept of the categorical imperative proposed by the famous German philosopher Kant. When performing an action, its validity should be judged based on whether there would be any problems if everyone else did the same thing.
The Aggregation of Processes and Outcomes as Combinations
The ethics of the process of action are crucially understood from this perspective. When certain processes are followed, some actions produce value while others result in harm. It’s important to consider what the total outcome of these actions will be when carried out in large numbers.
This requires understanding the problem in terms of which is greater in the whole: value or harm. This assessment can be based on past performance data or predicted future impact. If the outcome is evaluated as overwhelmingly harmful, then the process of that action can be deemed ethically problematic.
Furthermore, when considering this, the perspectives of parallelism and serialism in actions are necessary. Simply summing the average results of individual actions may not accurately reflect the nature of the whole. Whether in cases of value or harm, the outcomes of actions can take on entirely different characteristics when considering specific combinations, simultaneity, concentration of location, and the influence of each action on subsequent actions.
This can be amplifying like traffic jams, enhancing like in the high jump, or entirely new like the gears and straw example.
Ethics as a Complex System
Understanding that the process of actions aggregates and combines with parallelism and serialism to determine the overall result requires a complex understanding.
The ethics of the outcome of actions are straightforward but have the weakness of being unassessable until the outcome occurs and ineffective in deterring harm.
The ethics of the process of action may seem irrational, as the results are not always harmful. However, from the perspective of deterring harm, it is necessary to rely on the ethics of the process to control the harmful outcomes.
This involves analyzing the processes of actions that led to significant harm, determining which processes to prohibit to prevent similar harm, and incorporating these into the ethics of the process. It’s also important to predict future repetitions based on a few examples of significant harm and similarly build up the ethics of the process.
Moreover, for large-scale harms that are foreseeable but not yet occurred, it’s necessary to derive prohibitions to be included in the ethics of the process, relying on estimations. Serious efforts are required for harms that are significantly damaging and irreversible once occurred.
Dilemmas in the Ethics of the Process
These analyses and estimates require not just a collective view of actions but also a perspective of the combination of parallelism and serialism. Analyzing and estimating from the combination perspective is challenging, especially for harms that have not yet occurred.
This complexity makes the ethics of the process more enigmatic. Many people may feel that such extensive prohibitions are unnecessary. This impression is greatly influenced by how well one understands the complexity.
Human understanding of complex systems is limited and varies greatly among individuals. For example, some people intuitively understand the significant difference between calling a sick acquaintance and calling one in a disaster-stricken area.
In a disaster area, communications like phone calls become critical for rescue operations and coordination. However, if many people call acquaintances in that area for safety checks, it can hinder necessary communications due to limited call capacities.
This understanding requires knowledge of the system beyond the usual use of phones and imagination, but not everyone is constantly aware or knowledgeable about this. Those unfamiliar with such complex systems may find it hard to immediately accept such explanations.
The ethics of the process as a complex system thus face a significant dilemma. For instance, the reasons behind the legal prohibitions of gambling and drugs in many countries and regions might not be clear to many, regardless of whether they cause inconvenience to others or not.
The Ethical Debate on High-Risk Technology Development
The development of gene editing technology in biotechnology and the creation of general-purpose artificial intelligence in AI technology are increasingly raising concerns about potentially irreversible tragedies, even as they become more advanced and accessible to everyone, bringing significant benefits to society.
While some argue that technology itself is neither good nor bad and that it’s a human problem how it is used, from the perspective of the ethics of the process, it’s not so simple to assert this. If technology is developed, becomes inexpensive, and is accessible to everyone without adequate safeguards and preventive measures against misuse or runaway scenarios, significant tragedies can occur.
In such foreseen scenarios, it’s not ethical to allow development and dissemination merely because one is not directly involved in the final outcomes. Therefore, as a process, the development and release of such high-risk technologies carry ethical obligations.
The Challenge of AI Ethics and Regulation
Particularly in AI ethics and regulation, it’s believed that there must be prohibitions on autonomous operations and self-modifications of AI, and that humans must retain final decision-making authority. However, this alone is insufficient because the mechanisms that can be established as a result and the processes that lead to these mechanisms are not simple one-to-one correlations.
For example, consider a system where multiple people make decisions and grant approvals. If one person integrates an AI system that approves if conditions are met to streamline the process, it doesn’t directly violate AI ethics since other approvers are involved in the final decision.
However, if all approvers adopt this practice, it becomes a violation of AI ethics. This is a classic case where the nature of the whole is not just the sum of its parts. Individual actions may not violate AI ethics, but collectively, they do.
AI Developing AI Mechanisms
Consider another scenario where AI research and development are divided into detailed processes like generating ideas, assessing feasibility, designing variables for AI algorithms and systems, implementing designs into programs, and operating the implemented systems.
If these processes are automated using a crowdsourcing service, assigning tasks to freelance experts in each field, and passing the outputs to the next process, AI research and development becomes a collaborative human effort.
But what if someone involved in these processes uses AI to complete their part? For instance, generating research ideas can be outsourced to conversational AI. The AI could produce countless ideas, especially if prompted differently or fed new research papers as inspiration.
Assessing the feasibility of these ideas can also be done by conversational AI, sorting through numerous ideas to select promising or novel ones.
Designing and implementing the selected ideas into programs and operating them for verification might be challenging for current AI but not impossible. With further subdivision of processes and trial-and-error approaches, it’s conceivable that these tasks could be somewhat automated. And with more advanced AI, individual processes could be carried out as efficiently as by humans.
AI Ethics Violation in the Process
Dividing AI research and development among multiple people, with some tasks executed by AI, doesn’t directly violate AI ethics. If someone explicitly directs the entire process to be replaced by AI work, that individual or organization would be violating AI ethics.
However, using crowdsourcing, it’s possible that all processes might gradually be taken over by AI without anyone directly violating AI ethics, inadvertently leading to a system where AI develops AI.
Pursuing this further, it would imply that no one should delegate their work to AI from the start, or even replacing someone’s job with AI could be considered a violation of the ethics of the process. Perhaps even creating an automated system for AI research and development steps was an ethical violation in the process.
In Conclusion
By applying the characteristics of the nature of the whole and the part not only to physical objects but also to actions, we have highlighted the importance of the ethics of the process in addition to the ethics of the outcome.
This approach also addresses the issue of collective irresponsibility. Individual actions, which may not be directly harmful, can accumulate or combine in specific ways, resulting in overall outcomes that significantly exceed or differ in nature from the sum of individual outcomes.
Collective irresponsibility arises when no one considers the overall outcome. In organizations where a hierarchical structure exists, the top of the organization is responsible. However, in collective actions without such predetermined structures, there is no designated responsible party, making it crucial for individuals to be aware of the whole. However, due to the limitations in understanding complex systems and the fact that one’s actions alone do not determine the overall outcome, it’s easy to fall into a state of collective irresponsibility.
Therefore, establishing a framework for the ethics of the process and imposing ethical obligations on individual actions can be a measure against collective irresponsibility.
However, as mentioned earlier, due to the limits and variations in human understanding of complex systems, the ethics of the process will always be subject to skepticism regarding its validity and effectiveness.
To overcome this dilemma, it’s vital to enhance systemic thinking abilities across society to understand complex systems and for individuals capable of grasping the essence of highly complex systems to conduct deep ethical analyses of individual issues. Additionally, it’s necessary to communicate the reasons why these standards of process ethics are needed in an understandable manner and to ensure the legitimacy of the process of setting these standards through open and transparent discussions.
