Summary of Key Points
The Pope of Rome has issued the first AI ethics encyclical in the history of the Catholic Church, "Great Humanity," which focuses on the ethical challenges of the AI era and calls for the preservation of human responsibility and the prevention of the concentration of technological power. Meanwhile, Nobel Prize-winning physicist Pierre Parisi was interviewed, addressing key issues from a scientific perspective, such as the non-neutrality of AI, its black-box nature, human decision-making authority, the impact on employment, and global governance. He emphasized that AI should be a tool, not an authority, and that it requires multi-stakeholder participation and international regulatory frameworks.
Detailed Analysis
1. AI is not a "neutral tool"; it embodies the values of its creators
Pierre Parisi stated that "technology is never neutral," especially in the case of AI. For instance, the samples used to train AI algorithms (data) are not objective. If historical data used for recruitment show that more men were hired, the algorithm may learn and even amplify this bias (for example, by preferring to reject female resumes). The choices made by the creators (what data to use and how to design the model) inherently embed values. He advised scientists not to pretend to be "absolutely neutral" and to actively decide what values AI should uphold, such as fairness and respect.
2. AI is like a "black box," but it is not completely incomprehensible; future researchers may need to become "AI psychologists"
The deep neural networks of AI are essentially black boxes—we know what the input and output are, but we cannot explain how the intermediate processes work (for example, how an AI identifies cats from images). This is as unpredictable as complex systems studied by physicists (such as weather patterns). However, Parisi believes that it is not impossible to understand these processes. Physics provides tools, and in the future, new professions may emerge, similar to psychologists studying human behavior, to understand AI by observing its responses to different inputs rather than relying on mathematical equations.
3. Humans must retain ultimate decision-making authority; machines cannot make moral choices for us
The Pope emphasized that "the final decision rests with humans," and Parisi fully agrees. Machines can assist with data processing and provide recommendations, but they cannot make responsible decisions because they lack consciousness, understanding of right and wrong, and empathy. For example, AI should not be used to decide who qualifies for loans (which could discriminate against the poor) or who should be hired (due to bias), nor should it determine targets in warfare (the encyclical also opposes this). The most hidden risk is to reframe political or moral decisions as "technical issues" and let machines take the blame.
4. AI is transforming employment and power; the pace of change is too fast for us to adapt, and those who control the "digital gateways" will wield significant power
The impact of AI on employment is faster than that of the Industrial Revolution—changes used to take generations, but now they occur in just a few years, leaving people with insufficient time to adapt. More importantly, AI may become the gateway to our digital world (the information we see, the courses we take, and the jobs we get are all filtered by AI). Companies or organizations that control these gateways will have substantial power. Parisi believes this is not a fault of technology but a matter of how society distributes the benefits; for example, by reducing working hours to ensure everyone has access to opportunities, rather than a few individuals monopolizing all advantages.
5. AI requires global regulation, along with the involvement of philosophers and public research
Parisi believes that AI is not beyond governance (just as urban traffic can be managed with traffic lights). However, there is currently a lack of international regulations. He supports the establishment of public AI research institutions, similar to CERN for nuclear physics, because cutting-edge AI research is often held by large corporations that may keep information confidential. Public institutions can promote knowledge sharing and encourage participation from smaller companies. Additionally, AI issues go beyond technology; questions such as "what constitutes consciousness" and "who should be responsible" have been debated by philosophers for centuries and cannot be resolved by engineers alone. Therefore, philosophers, humanities scholars, and ethical organizations need to be involved in the discussion. He also calls for a reorganization of the international governance system (e.g., giving the United Nations more authority) and even the signing of international treaties to establish clear boundaries (such as prohibiting AI from making autonomous decisions that could result in harm).
Conclusion
The core message of this interview and encyclical is that AI is a double-edged sword. Its benefits and drawbacks depend on how humans design, use, and regulate it. We must prevent technology from getting out of control and prioritize human dignity and responsibility. This will require global cooperation and the collective wisdom of various stakeholders to address these challenges effectively.