Moral Leadership in a Time of Great Need

Artificial intelligence (AI) holds the potential to address the world’s most difficult problems. From accelerating drug discovery and extending access to quality education in underserved communities, to improving energy management and enabling earlier diagnosis of disease. Pope Leo XIV's first encyclical, Magnifica Humanitas, released on May 25th recognizes that potential, while making clear that promise alone is not a sufficient basis for policy. The Pope calls on governments to act across two broad categories: 1) development of a more robust social safety net that can support workers and communities as AI reshapes labor markets and further concentrates economic power, and 2)stronger governance of AI development and deployment, establishing clear rules about how the technology may be used, by whom, and under what conditions. The goal of these rules and oversight is protecting those in society with the least power. These are not new ideas. They are the application of long-standing principles of social justice to a technological moment. It is critical they receive the seriousness that the Pope brings to them.

The concern about unemployment runs through Magnifica Humanitas. The encyclical does not treat job displacement as a distant or speculative risk. It treats it as a present condition that governments have an obligation to address. The encyclical calls for a renewal of labor organizations and for rethinking how economies measure progress, moving beyond GDP toward measures that capture the dignity of work, strengthened communities, shared prosperity, and the reduction of inequality. Governments that respond only through retraining programs, without addressing the conditions under which firms adopt AI and without strengthening the safety net beneath workers who lose their positions, will find the response inadequate to the scale of the disruption. 

We have seen this in our research at Duke on AI’s impact on the music industry. That research has convinced us that AI’s effect on musicians is a “canary in the coal mine” for what will happen to many sectors of the economy. The Pope makes this point clear, stating that governments must react to the likelihood that when businesses implement technology in pursuit of reducing costs and increasing profit, it will lead to mass unemployment. That framing locates the responsibility for this effect with both governments and with those making decisions about how organizations will deploy AI. The Pope's framing requires a shift for justice to an economy that values dignity, not just output.

The governance challenge extends beyond the loss of jobs. Anthropic’s Mythos model demonstrates the destabilizing societal risks if AI solutions are released without proper governance. Experts with access to the model have stated that it provides a step-change advancement in being able to find cybersecurity vulnerabilities in hardware and software. Anthropic’s voluntary decision not to release the model mitigated the risk of malicious actors having access to these vulnerabilities. If not for their voluntary decision, technology companies around the world would have to race to develop fixes and distribute patches in lack of oversight frameworks. That is a race those companies clearly would lose. Malicious actors could compromise critical infrastructure such as the power grid and water treatment plants. Depending on companies to voluntarily protect the public from these types of risks is a fundamental failure of governments in their primary role to protect their citizens.

The encyclical's warning that control of AI must not remain in the hands of a few is a direct response to this kind of situation. When the safety profile of a frontier model is only evaluated through processes that the company designs and acts upon only through voluntary decisions by that company's leadership, the governance structure is not adequate for the risks at stake. Societal safety should not depend on decision making of unelected company CEOs. The question of whether a model like Mythos is released to the public, or deployed in sensitive contexts, should not rest on a voluntary decision by any single company, however well-intentioned. The Pope is right to insist on something exponentially more responsible. 

The potential for the use of AI in warfare makes clear why mandatory and robust governance is required. The encyclical condemns in categorical terms the use of lethal autonomous weapons systems that operate without meaningful human oversight. While some argue that AI systems have lower error rates than humans in targeting decisions, this technical capability misses a fundamental moral point. Pope Leo XIV described these risks in an address at Rome's La Sapienza University days before the encyclical's release, pointing to what is happening in Ukraine, Gaza, Lebanon, and Iran as evidence of what he described as an “inhuman evolution of the relationship between war and new technologies in a spiral of annihilation.” In the speech, he continued the call for governance so that AI “does not absolve humans of responsibility for their choices and does not exacerbate the tragedy of conflicts.” Even if AI reduces technical errors, delegating life-and-death decisions to machines removes human moral accountability and creates a gap between the decision-maker and the harm. Discussions are immediately necessary between governments to set norms and binding accountability structures so that AI is not used in fully autonomous weapons.

The encyclical also calls out the risks to freedom from mass surveillance in the following passage:

A further risk, less visible but no less serious, is that of social control made possible by the massive collection of data and use of algorithmic systems. When every action—movements, purchases, relationships and preferences—leaves a trace, a new form of power emerges, namely the power to profile, predict and influence behavior, often without individuals being fully aware of it. If such kinds of data are used to make decisions affecting concrete opportunities — such as access to credit, employment or essential services — there is a risk of undermining freedom and discriminating against the most vulnerable. Furthermore, control is exercised not only through explicit prohibitions, but also through the architecture of visibility: what is amplified or rendered invisible, what is rewarded or penalized, ultimately shapes opinions and choices, fostering conformity and self-censorship. For this reason, freedom in the digital age is not merely a matter of interiority but also a public concern. It calls for clear rules, transparency, the possibility of recourse and proportionate limits on the use of intrusive technologies, so that technology will remain at the service of the human person and not become a form of control over consciences. (Paragraph 171)

The risk of use of AI for mass surveillance is one that the developers of all frontier models should recognize. Both governments and private sector companies should not be able to use AI to further mass surveillance as the technology will then further erode the dignity of the individual. Dignity requires a private space to make decisions, learn, and exercise civil liberties. Preventing these risks has been at the center of our Duke work on data brokers, platform accountability, and privacy legislation. I would extend the encyclical’s comments on this topic to note that enforceable privacy protections for the individual, with access to the courts, are critical so that all people can move freely, love who they choose, seek reproductive healthcare, and be free of the threat of stalking and other acts of domestic violence. AI implemented poorly could further erode all of those human rights. In the absence of a strong comprehensive federal privacy law (something I have advocated in support of for more than twenty years), it is critical that U.S. states continue to strengthen their privacy laws and increase enforcement.

Our work at Duke has at its foundation analysis of how technology impacts the least powerful in society. That approach is grounded in the rich history of the Catholic church’s liberation theology, specifically the writings of Gustavo Gutierrez and James Cone. To take into consideration AI’s impact on the least powerful, we should analyze the reach of AI's consequences beyond human communities. The encyclical's moral framework, grounded in the protection of dignity, extends to sentient non-humans. The documentary AI and Animals, produced by the nonprofit Animal Ethics and released in early 2026, examines this dimension directly. The film brings together philosophers, scientists, and advocates to describe what AI could mean for nonhuman animals, who have no voice in the current governance debates. The potential for benefits and risks are both substantial. These benefits and risks are indicative of the potential for both good and bad results more generally from deployment of AI if more is not done in model development to maximize the positives while minimizing the negative impact. 

AI can help monitor disease to animals, point people to plant-based protein, protect wildlife, and decode animal communication. All of those AI uses could benefit the lives of trillions of animals over time. However, AI-driven optimization in industrial agriculture could also accelerate what the documentary describes as factory farming 2.0. That new generation of farming could in the interest of efficiency and profit enable higher densities, faster production cycles, profiled marketing of animal-based food, and reduced human oversight of animal welfare. The documentary's core argument is that the values embedded in AI systems now will shape outcomes for animals for decades, and that the window to make different choices is closing. Moral leadership that is serious about the common good cannot stop at the boundaries of the human species.

AI and Animals closes on a note that resonates with the argument of Magnifica Humanitas. We are spending billions of dollars on AI innovation. To allow society to reap the benefits of that investment we must spend a similar amount on robust governance and enforcement. Alignment to pursuit of the common good does not happen through corporate voluntary commitments or industry self-regulation alone, however sincere. It happens through the kind of moral and institutional leadership the Pope has demonstrated and in making demands of governments that are specific enough to act on. It happens because government officials exercise leadership and courage to put in place strong rules, robust resources, and accountability demanding enforcement to protect society. 

The current polling data on the public’s distrust of AI speaks with a loud voice to elected representatives. It is time for that voice to matter more than the corporate donations from technology companies, investors, and their executives. The public policy society needs must rely upon bringing together technologists, social scientists, civil society, and government to engage the questions seriously. We need to bring both science and the humanities together to reduce these risks. This is the type of interdisciplinary research and approach to public policy that forms the foundation of everything we do at Duke. Our work points toward the same conclusions as the Pope’s encyclical that the ethical use of technology requires governance structures with real legal powers and resources, not only principles and aspirations. Magnifica Humanitas makes that case with moral authority. The response it deserves is not just admiration. It is action.

David A. Hoffman is the Steed Family Professor of the Practice of Cybersecurity Policy at Duke University and is the Interim Director of the Duke Initiative for Science & Society