Last week, the Bioethics Observatory of the Catholic University of Valencia held the conference “Humanity and Algorithms: Bioethical Reflections on AI.”
The meeting was inaugurated by José Manuel Pagán, rector of the UCV, who invited attendees to “develop a critical perspective on artificial intelligence.” To this end, he urged that “ethics should always be present in this debate”. He is convinced that it is “the only guarantee of authentic progress in the service of the common good.”
He also emphasized that “the university must form consciences capable of discernment” and stressed that “technology must always be at the service of the person, not the other way around”.
During the opening ceremony, Julio Tudela, director of the Bioethics Observatory, invited participants to “present the major ethical challenges posed by artificial intelligence and underline the urgent need to address them from a bioethical perspective.”
He also called for “ethical analysis to advance at the same pace as technologies, to prevent risks from outweighing opportunities.” He also called for “defending the freedom, truthfulness, and value of human intelligence against possible manipulation and dehumanizing uses of AI.”
Eclipse of reason
For Cortina, communicative reason is important because it is intersubjective, as the German philosopher and sociologist Habermas explains. “The creation of the ethics of dialogue supports the creation of agreements on common goals,” she noted.
Cortina clarified that when we talk about AI, we should actually talk about it “as if it were intelligent,” that is, the “as if” should be used before the term “intelligent,” since in reality it is not.
The professor spoke of a catastrophic vision of artificial intelligence that opposes AI enthusiasts and proposes ethically educating machines according to moral values. In both cases, both the catastrophic vision and the enthusiastic vision, they need to be built on ethical criteria.
For Cortina, there are different ethics applicable to this phenomenon, such as the ethics of superintelligence, the ethics of general intelligence, and the ethics of special intelligence. Human intelligence appeals to common sense, while instrumental intelligence focuses on specific topics. The other ethics are still out of our reach.
According to the expert, in Europe there is an ethical framework for artificial intelligence that uses the premises of principlism. To this, she added other necessary principles: beneficence, non-maleficence, autonomy and justice. She also added the principles of explainability, accountability, and protection of nature.
Finally, she insisted that when we talk about autonomy, we should be talking about functional autonomy, because artificial intelligence is not autonomous.
Algorithms and AI
Enrique Estellés, professor of Information Technology at the UCV, based his presentation on the structure of algorithms. To understand how AI works, he asserted that it is necessary to understand the mechanisms underlying these tools.
Estellés explained that algorithms, the essential core of artificial intelligence, are designed to mimic, as closely as possible, the neuronal functioning of the human brain. He also noted that these algorithms are based on data collection and analysis, so the quality of that data will be crucial to the reliability of the results they provide in our searches.
Privacy and security in personal data
“It is essential,” Martínez emphasized, “to protect the well-being of the community, to improve diagnostic tools to establish the limits of artificial intelligence tools and their risks, as well as their potential.
It is also necessary to establish a governance framework that controls the system, for example, by setting limits and prohibitions on intentional harm, preventing these tools from intentionally setting out to cause harm.
Diversity in data collection, cross-borderization, that is, globalization in data management, can be both a tool and a risk. Protecting the privacy and security of personal data is essential.
The expert emphasized the need to conduct audits to analyze procedures and their results, security, veracity, and reliability. Legal regulation is also essential, which must focus on the human being and promote AI literacy processes.
Doctor-patient contact
Federico de Montalvo, professor at the Comillas Pontifical University-ICADE, focused on the doctor-patient relationship, defending the corporeality and human contact between the two: “personal contact is irreplaceable,” he asserted.
De Montalvo stressed the importance of silences, something that artificial intelligence does not understand, and sometimes silences are as eloquent or more eloquent than words. These can only occur in personal relationships, and a diagnosis cannot be reduced to an impersonal algorithm, although it can improve it, but it has to be linked to a narrative in which the physician plays a fundamental role.
Autonomy versus dignity
Modernity has prioritized autonomy over dignity; that is, one of the tendencies of modernity is to give more importance to personal autonomy than to human dignity.
The European Union’s regulation on AI omits the term dignity; however, a previous UNESCO recommendation does mention the term dignity. It analyzes the potential risks these tools may pose to respect for human dignity.
The professor recalled that the Vatican’s note Antiqua et Nova stresses that common sense prioritizes dignity over autonomy, promoting humanization in the use of these tools.
Requirements instead of principles
The director of the Chair of Bioethics at Comillas Pontifical University, Rafael Amo, urged us to define the epistemology of artificial intelligence, that is, its nature, its purposes, and its usefulness.
The bioethicist spoke about the ethical criteria that must be followed for the use of AI, including ethical artificial intelligence, that is, attempts to provide artificial intelligence itself with ethical criteria.
He stated that work is currently underway on this: “Rather than talking about bioethical principles, we must first establish an ethical framework for AI, assessing the impact on people and the environment. It’s more appropriate to talk about requirements than principles.
Among these requirements, which would build algorithms, we should remember some such as explainability, autonomy, privacy, robustness, fairness, sustainability, wellfare and accountability. All of this would have to be promoted through artificial intelligence”.
Physician autonomy, threatened by AI
He noted that principlism, as the dominant current in bioethics, is permeating the ethical frameworks of AI, which can lead to a reductionist view of medical ethics.
He also warned of the danger of excessive dependence on AI, even on the part of ethics committees, which could end up delegating part of their work to these systems. Finally, he stressed that the ethical dimension of the doctor-patient relationship, which is essential for the full and humane practice of medicine, is being undervalued.
Preservation of confidentiality
José María Domínguez, president of the Medical Ethics and Deontology Commission of the Spanish Medical Association, presented the recently developed Declaration on the European Health Data Space.
Domínguez explained to the attendees the precautions the General Council of Physicians observes in the European Community’s documents for the ethical regulation of AI. “The preservation of confidentiality in data handling does not seem to be guaranteed, so physicians must demand greater rigor in data handling when there is no mandatory authorization from the patients.”
“Real” versus “virtual”
Ciucci was optimistic about artificial intelligence tools. In this sense, he stressed that technological development contributes to truly human development, and therefore, we must guide Artificial Intelligence and learn to use it in the digital era.
The expert also criticized the division between “real” and “virtual,” recalling that virtual is also part of reality, so such separating boundaries should not be established.
This was later discussed by Alfonso Ballesteros, Professor at Miguel Hernández University, and Vicente Bellver, Professor at the University of Valencia, who led the final round table.
Humanity and algorithms
Bellver expressed his concern about the speed of implementation of artificial intelligence, although he was not opposed to new technologies. “This is happening in the face of an uncritical attitude on the part of the population, and this is very dangerous because it can lead to very destructive processes such as loneliness, isolation, and the pursuit of self-satisfaction. It is necessary to ask ourselves whether or not AI contributes to true progress, to human progress.”
Ballesteros was more cautious in his view of AI compared to the optimism expressed by Ciucci, warning of hyperinnovation, that is, “it seems that innovation is already a value in itself and not all innovation represents progress, therefore, we have to be cautious. We need to set aside time to think.”
In his opinion, the globalization of data creates social and political polarization and promotes loneliness. The use of chatbots, software programs designed to simulate a conversation with a person, as substitutes for personal relationships is proliferating. This is bad news because they can promote authoritarianism and social disorganization.