EU Bishops highlight ethical foundations of AI

By Fr. Benedict Mayaki, SJ

The Commission of European Bishops (COMECE) has called on the European Union (EU) to adopt a “human-centric approach to Artificial Intelligence (AI) in order to promote the common good and serve the lives of all human beings in both their personal and community dimensions.”

The Bishops made this call as part of their contribution to a recent publication by the EU titled “Consultation on the White Paper on Artificial Intelligence – A European Approach.” 

Humans, different from AI

Welcoming the general approach of the White Paper (authoritative document intended to give information on a subject) to establish AI that is deeply grounded in human dignity and protection of privacy, the Bishops highlighted that data and algorithms are the main drivers of AI, while humans “determine and program the goals which the AI system should attain.”

COMECE called for a clarification of the term “Artificial Intelligence”, adding that it be “clearly differentiated from human conduct.” This is because the Christian perspective sees the human person as “qualitatively different from other beings, with a transcendental dignity, intelligent and free and capable, therefore, of moral acts.”

“AI systems are not free in the sense the human person is and, in this sense, its act cannot be judged according to the moral criteria that are applied to human acts,” the Bishops stated.

European governance

The European Bishops said that they were “perplexed” about the possible creation of a new EU agency to regulate AI and robotics, adding that “it is important to curb excessive multiplication of Union structures.”

However, in the case that such a body is formed, the Bishops advocate for a governance structure that “guarantees maximum stakeholders’ participation,” including churches, which “have a specific status as partners in the EU institutions.” 

AI against money-laundering

According to the Bishops, it is important that the use of AI in the fight against financial crimes – especially money laundering – does not lead to a “society of control” and “undue interference in the organization of churches and charitable organizations.”

“It is crucial to balance transparency with privacy and autonomy,” the Bishops said, adding that this is especially true in the context “of an increased recourse to AI in countering these phenomena.”

Military systems

While indicating that the White Paper did not touch on military purposes of AI, the Bishops nonetheless urge the EU to “ban completely autonomous armed systems without human supervision” and work towards “international negotiations on a legally binding instrument prohibiting lethal autonomous weapon systems.” 

Cyber security

The Bishops pointed out that while AI may enhance security in a digital environment, it can also open up new vulnerabilities. They gave the example of digital diplomacy where the use of AI can have far-reaching consequences for the democratic order through the uncontrolled spread of disinformation, or external influence by non-state actors.

In light of this, they proposed the creation of mandatory requirements for risky AI technologies which can affect public safety. They also called for the strengthening of critical infrastructures against AI-induced security challenges, and scrutiny of private companies and beneficiaries of final control of the collection and analysis of personal data.

Fundamental rights concerns

The COMECE document also contains detailed proposals in relation to fundamental rights including liability, safety, algorithms, children and protection of personal data.

As regards to safety, the Bishops stressed the additional obligation of manufacturers to provide features that prevent the upload of software that impacts on safety during the lifetime of AI products. 

The Bishops also spoke up for the protection of children, pointing out that the child is “the most vulnerable actor in the context of AI use and application.” They insisted that manufacturers should have explicit obligations to consider the immaterial harm their products can cause to children, vulnerable users, and elderly persons in care environments.

The COMECE contribution also calls for the regulation of aggressive data collection and intrusion in citizens’ privacy, saying that while data is needed for AI to function, human control should remain at the center of AI use as this plays a role in “upholding high data protection standards.”

In February 2020, COMECE participated in the internal workshop “The Good Algorithm? Artificial Intelligence, Ethics, Law, Health” which was held at the Vatican on the occasion of the 26^th General Assembly of the Pontifical Academy for Life. On that occasion, the “Rome Call for AI Ethics” – a document stating the resolve that digital innovation and technological progress would be at the service of human genius and not replace it – was signed.