In this paper, we would like to propose an application of the concept of sustainability to technological processes and in particular to the elaboration of AI systems, for the construction of an ethics that could confer meaning, motivations and impetus in using new technologies and Artificial Intelligence. To reach this goal we will first try to build an ethical framework, here called Dynamical Techno-Algor-Ethical Composition, to define the interaction between the most important ethical ingredients involving the human person in relation to technology, taking a person-centered approach. Out of this it will emerge a possible structure and definition of Technological Sustainability. The second step will consist of evaluating the process for the elaboration of an AI algorithm as a concrete application of the previously analyzed framework, to set ethical contents composing what we might call a good and sustainable algorithm.
We first introduce the possible basis of attempting to set a motivational and operational framework for a wise interpretation of technological development, in order to orient it to the service of the human person and therefore to his good. It will then be possible to propose a definition of Technological Sustainability in the context of an adequate ethical person-centered reference, and to explore its application in the development of a consequent sustainable Algor-ethics applied to AI.
In the context of a Technology strictly connected with Anthropology, that is:
we first pose two important and precious pillars to structure these relations:
The latter inscribe, in our context, an orientation of meaning, giving the necessary breath to the Social Ethics of Technology, which otherwise would remain prisoners of a freedom without a source. These coordinates are not static but are strictly interconnected in the dynamism of human life so that the ethical attitude could emerge as a result of their synergy. Let us now explore their main traits:
Social Ethics of Technology:
Justice is the way through which the ethical commitment of charity takes substance, which is in turn embodied precisely in Justice, without, however, identifying itself with it; in fact: “Charity . . . has its first expression in Justice, which . . . informed by charity, participates in the theological and salvific tension” . It is in this sense that charity prevents justice from objectifying itself, that is, from remaining static and arid, just as, without justice, charity would be abstract and unproductive: “Justice gives charity a realistic incarnational character” .
Transcendental Motives for Technology:
Could be considered as three dynamic ‘engines’ that could orient technology itself in a fermentation process that, on one side evaluates human’s potentialities, and on the other maintains a look to human dignity and its fulfillment:
Receiving as “inputs” the Social Ethics of Technology (responsibility, justice and power) on the one hand and on the other the Transcendental Motives for Technology (anthropology, freedom-creativity and eschatology), we now propose a specific composition in order to appropriately structure them in a dynamic, operational and propulsive relationship: dynamic, because all the individual elements of such a system should be continually interconnected in an active dialogue, capable of wisely facing the complexity of reality and the challenges it poses; operational, as the integral relationships between the components should be able to descend to the level of implementation, and therefore have strong roots and luminous foundations; propulsive, because an organic reading, such as the one we are proposing, in addition to offering a solid elasticity, should finally be able to push towards a fullness of meaning and choices regarding the good of the human person.
With these premises, we now come to introduce two elements that we believe are decisive for guaranteeing the right glue and the necessary elasticity to all the analyzed elements, the horizontal and the vertical ones, in order to structure bonds with the characteristics described above: the Service of Love or Power of Service (first element), which is able to harmonize the structure we propose, precisely in the horizon of meaning that calls the human person to his Fullness (second element).
This is to obtain what we call the Dynamical Techno-Algor-Ethical Composition, which constitutes the ethical environment in which to develop specific traits and operational provisions in the individual technological and/or algorithmic design applications (Figure 1).
Figure 1. The Dynamical Techno-Algor-Ethical Composition.
Figure 2. (a). Systolic beat of Service of Love. (b). Diastolic beat of Service of Love. (c). Breath of Responsibility.
In this way the Social Ethics of Technology and the Transcendental Motives of Technology are dynamically connected and receive, as a living and enlarged heart, the Service of Love, which constitutes their ‘beating’ synthetic core, and also the Completion, which represents in turn their propulsive orientation term: all this defines the Dynamical Techno-Algor-Ethical Composition or Ethical Framework for technology finally composed with all these macro-ingredients:
This dynamic composition acts as an ethical reference for technology and AI, which should be inserted within it, and in this way could lead to the construction of a Sustainable Technology and Algor-ethics.
At this point we can summarize Technological Sustainability, offering this definition:
“Technological Sustainability is the opportunity to develop an authentic human progress according to an ethically consistent interaction with the technique, considered as a person-centered dynamical service. Technology, in this context, assumes its own role to sustain mankind’s freedom and creativity in order to respect human dignity and to facilitate, embracing with responsibility all humanity in the service of love, the expression of its common good in a diachronic path toward fulfillment”.
In this section we propose a specific application of the ethical framework that we have presented, to Artificial Intelligence. The Dynamical Techno-Algor-Ethical Composition could be in fact, in our proposal, the framework in which a researcher may move to build and structure a sustainable technology or Good Algorithm, being disposed in an ethically consistent perspective. For this purpose, we choose to split an AI project into 5 subsystems, following the different steps that realize its development:
An Algor-ethical monitoring in fact, should not be developed simply starting from the inputs and outputs, but should proceed by analyzing all the phases of the algorithmic process, thus entering the technological and algorithmic box to evaluate all the elaboration chain from the inside.
The aim of each following paragraph is to answer to the question: “which ethical trait could emerge from each single algorithmic step drawing on from the Dynamical Techno-Algor-Ethical Composition?”.
Data Ethics: Poverty of Spirit
The data represent a numerical sampling, therefore a translation into digital numbers of quantities associated with the dynamic and continuous life of a person, or with natural phenomena. It is the very first step that collects a sort of digitalization of man and reality, a form of sampling to provide an adequate and sufficient virtual image, which will then be artificially processed or elaborated.
It therefore seems appropriate to combine this precious moment of compiling and using an algorithm with the attitude of the ‘poverty of spirit’. This provision, on the one hand, should always be maintained throughout the evolution of the technological process (design, construction and use); on the other hand it allows us to interface with Big Data in an ethically sustainable way. In our context the ‘poverty of spirit’ consists in fact of a disposal that guides us to consider the data for what they really are, ‘respecting them’ within the poverty/totality polarity:
Model Ethics: Meekness and Minority
Faced with the importance and diffusion of models, we then propose to combine the attitude of ‘meekness’ at this stage of technological development. Meekness could be the disposition with which man builds the model, freeing it from the presumption of being in the likeness of man. At the same time, man could confer to the model the traits of meekness, as long as it remains in the ‘minority’ with respect to the reality.
The ‘meekness of the model’ is then its ‘technological minority’, a minority that does not mortify its power and subtlety, but which instead gives it back its strength at least in two directions: that of its increasingly advanced potential and that of the reference and return to man not only as an author, but also as a decision maker, evaluator and measurer.
Learning Ethics: Hunger and Thirst for the Truth
Machine Learning defines and tunes numerical parameters, structured in a more or less complex network of interconnections. These are numerical parameters which, as they are derived in an algorithmic way from Poor Big Data, in turn should be considered in the same way as Poor Parameters. The sequence is therefore as follows:
These parameters, which define an artificial learning model, always carry with them a more or less evident component of error, partly due to the quality of the data themselves and partly to the inevitable limitations of the model. Machine Learning also lacks, with respect to human learning, the dimensions of relationality and creativity, of conscience, emotions, feelings, judgments and previous lived experience, freedom, neural conformation and the uniqueness of the Person, in his unrepeatable unity. So, we could summarize by saying: “when faced with Machine Learning, ‘remember to learn’”, as an opportunity to feel always ‘hungry and thirsty’ for human learning and therefore for knowledge and truth. In fact, in the case of technology, knowledge tents to be always and only technical, reaching further data, while in the human case knowledge is always oriented towards the relationship:
Validation Ethics: Pacification
In this part of the journey in the technological algorithm, we recognize the greater need for ‘peace’ between the various competing fronts: the expectations of the designers; the quality of the input data; the quality of the model; the error of the results (standard deviation); the expectations and satisfaction of users.
This type of ‘peace operation’ can be defined as fine tuning; it is in fact a question of tuning different parameters, different processing qualities and different weights that can also conflict with each other, requiring design compromises.
We then see the validation of the data as a real peace-making process, where the most essential pacification is that between what is typical of human evaluation and what is numerical evaluation. In fact, human evaluation contains elements which, even if they may not always lead to absolute certainty, refer to an integral experience that cannot be reproduced and cannot be codified.
The pacification we are talking about therefore refers precisely to this ethical dimension of evaluation, which cannot in any way be attributed to the machine and algorithm, which are lacking of the essential components of the evaluating process. In the artificial chain of evaluation there is no imagination, intellectual intuition or judgment, which in the artificial does not have a predicative function (as for the Human Person) but of classification. It does not in fact reach being, but it reaches a statistical, numerical, qualitative, distributive, organizational and precisely classificative content. It is a more or less complex manipulation of data, not a knowledge or a judgment.
The first ‘pacification’ consists therefore of not speaking of Artificial Intelligence but of Artificial Processing or Elaboration, and the second in not referring to Machine Sapiens but always to Homo Sapiens also if enhanced with the aid of artificial instruments. The machine can never be wise, while the human person partakes of a personal wisdom. The synergistic and ethically balanced complex of human and artificial can instead be wisely considered by virtue of the human component alone:
Pacifying the moment of evaluation is therefore an essential step, because it places the truth of relationships and differences, restoring the dignity of knowledge to the human person. From this pacification derives the ethical evaluation of the interaction (not relationship!) with the machine and with the results it can produce.
Presentation Ethics: Purity of Heart
The final dimension of our algorithmic process consists in the phase of data presentation, that is, the interface with the user. It is a very important system as it involves the dialogic-communicative aspect. The impact of ‘presentation’, in this digital age has strong resonances on the inside of the persons. In fact, very deep chords of the human person are touched, precisely in his interiority and spirituality, and it is therefore important that the Algor-ethics appeals to a ‘pure’ disposition of the person who designs, and that defines the ‘dialogic’ dimension man-machine. This could be a fruitful and sincere way to present a great expression of human intelligence, which is technology, by placing it in its appropriate context of service to humanity and also of strengthening its faculties. The results presented with the attitude of the ‘pure of heart’ will then offer the user the opportunity to see the person and not just the machine, and even himself, perceiving the expansion to the mystery that, in the humble and non-imposing attitude of technology, directs him to open the doors of his interiority and spirituality, to become more and more himself, not remaining just in the technological level. In this way, technology could really serve the integral human person and enhance his faculties, allowing the gifts that characterize him to dare where humanity alone could not reach.
The set of these five ethical relationships, that now we resume as:
provides a paradigm that is not only exploratory but declarative of a process of re-reading technology in an ethical key, primarily involving the human person in the rediscovery of his dignity and his transcendence not available for reductionism.
Poverty, minority and meekness, relationality, pacification and purity of heart, are characteristics of the Service of Love emerging from an AI system designed within the Dynamical Techno-Algor-Ethical Composition. In fact, in these characteristics converge the ethical tensions typical of technology, such as power, justice and responsibility (Social Ethics of Technology), which are always lurking as risks and at the same time are always available as opportunities for growth along the entire path of a technology oriented towards the common good and fulfillment (Transcendental Motives of Technology). These features agree with the definition of Technological Sustainability and therefore define a sustainable Algor-ethics for AI.
In this way, both the technological process and, above all, the researcher are involved in an attractive and challenging dynamism:
The difficulties in applying this approach, which is based, as we have seen, on a multi-dimensional and dynamic ethical composition, for the various technological processes, allowing concrete ethical paths to emerge, mainly reside in the required change of mentality. On the one hand, the latter is in some way imposed by the conspicuous overcoming of mechanistic determinism; on the other, however, in the face of ‘quantity’ and ‘speed’ (two characteristic traits of the post-modern technological era) it risks being practically forgotten due to lack of time and therefore attention. This change of mentality introduces an ethical horizon as well as a transcendent horizon, as not only necessary but also propulsive resources.
Who knows if a new scientific beatitude will not be born from this!
This entry is adapted from 10.3390/su14063215