The AI ethics debate looks different from Manila

(RNS) — At my daughter’s recent graduation from a liberal arts college in the Northeast, artificial intelligence was repeatedly booed when it appeared in commencement speeches. The reaction did not strike me as simple technophobia. It sounded like moral apprehension.

My own young adult children voice many of the same concerns that I hear from colleagues in higher education from across North America and Europe: AI may weaken critical thinking, devalue human creativity, consume staggering quantities of energy and water, exploit poorly paid labor and accelerate a culture already too eager to exchange wisdom for convenience. Likewise, as I help develop AI resources for colleagues in theological education through a Wabash Center grant, the objections are palpable. Faculty worry that students using AI will replace their ability to learn to think. Scholars worry that their work may be absorbed without consent. People of conscience wonder whether refusing these systems is the only ethical response.

Then I came to Manila. I am here keynoting at the international triennial conference of the Colleges and Universities of the Anglican Communion, where leaders, drawn from over 150 institutions around the world, have gathered to consider the future of Anglican higher education in an age of AI. In preparation, we surveyed 30 leaders from 12 countries about how their institutions are approaching AI.

The results revealed a striking difference in emphasis. North American respondents were especially concerned about environmental impact and the possibility that AI might threaten the nature of higher education itself. Respondents from the Global South more often emphasized cost, infrastructure and equitable access. Twenty-four of the 30 institutions already use AI in teaching and learning, and 20 encourage students to use it with guidance.

This is a small survey, and “Global North” and “Global South” are blunt categories. Nearly half of the respondents came from India. Still, the pattern in the survey has been echoed repeatedly in conversations around tables, in lecture halls and over meals.

Colleagues have told story after story of institutions integrating AI not as a replacement for education but as a means of extending it. Rikkyo University in Japan has established a graduate school devoted to AI while requiring its students to study AI ethics. At Women’s Christian College in Chennai, India, colleagues described students using AI to help make educational materials available to language communities that might otherwise struggle to access them.

These institutions are not ignoring critical thinking. They are teaching students to think critically with AI rather than allowing AI to think in their place.

A colleague from Kenya made the point most memorably. I paraphrase: What looks to one person like a shortcut may be the only path available to another.

That sentence has stayed with me because so much English-language anxiety about AI comes from institutions that have benefited from generations of accumulated wealth, privilege and access — often acquired through the very processes of colonization whose legacies we are seeking to confront and repair at this gathering.

We have therefore come to imagine, perhaps without fully realizing it, that a “real” university requires an enormous library, a substantial endowment, layers of specialized staff, abundant faculty time and reliable access to every scholarly resource. These things are genuinely valuable. But for many institutions, they are also the products of histories in which resources and opportunities have been distributed profoundly unequally.

AI will not erase those inequalities. Indeed, it may reproduce and even deepen them through biased systems, expensive subscriptions and unequal access to computing infrastructure. But it may also reduce some of the distance between a wealthy university and one that has never possessed those inherited advantages.

Translation tools can open educational materials across languages. AI-assisted tutoring can provide support where student-to-faculty ratios are high. Faculty can adapt resources for students with different levels of preparation. Small institutions can extend administrative and research capacities that larger institutions take for granted. And students who are taught to use AI critically can learn to identify, interrogate and mitigate the biases embedded within these systems, rather than simply receiving their outputs as neutral or authoritative.

In this light, what constitutes an effective university may depend less on the size of its endowment or the number of volumes in its library and more on how creatively and responsibly its faculty use the tools available to support learning and formation.

There is a danger, then, that principled resistance to AI can become a luxury good. Those of us who already possess abundant educational resources may decide that the morally pure response is to close the laptop. Meanwhile, institutions serving students who have been historically excluded from those resources may see in the same machine or LLM a possibility for access, translation and participation.

That does not make every use of AI good. It means the moral question cannot be answered from the standpoint of inherited abundance alone.

People of faith (and dare I say, Anglicans in particular) must insist on a genuinely global ethics of AI. We cannot ignore the extraction of rare-earth minerals, the use of water and electricity, the hidden human labor behind automated systems or the wholesale appropriation of human learning and creativity. Stewardship requires us to confront those costs honestly.

But honesty also requires consistency. AI is not the only technology implicated in extractive systems. Smartphones, cloud storage, streaming entertainment and the devices on which many of us denounce AI share almost all of the same supply chains and environmental burdens. That does not excuse AI’s costs. It should, however, make us suspicious of moral arguments that isolate one technology while leaving our larger habits of consumption untouched.

In our survey of delegates to the conference, when asked to identify which values should shape Anglican higher education in the age of AI, the first was a concern for human dignity, followed closely by a call for wise stewardship. Those values belong together. Stewardship asks what these systems consume and whom they exploit. Human dignity asks who may gain a voice, an education or an opportunity that was previously denied.

As I have written elsewhere, a Christian response to AI should therefore resist both technophilia and technophobia. It should not only ask, “What does this machine do to us?” but also, “For whom might this machine make something possible?”

At my daughter’s graduation, the booing sounded like conscience. In Manila, the stories I am hearing sound like hope. We need both voices at the same table.

The deepest question is not whether we are for or against artificial intelligence. It is whether we can discern its dangers and possibilities from beyond the narrow horizon of our own advantage.

(The Rev. Michael W. DeLashmutt is dean of the Chapel of the Good Shepherd and senior vice president at the General Theological Seminary in New York City, where he also serves as associate professor of theology. His most recent book is “A Lived Theology of Everyday Life.” The views expressed in this commentary do not necessarily reflect those of Religion News Service.)