A day after Pope Leo XIV delivered a major declaration in Rome about the future of a world with artificial intelligence, a Latter-day Saint apostle at an important summit in Athens, Greece, shared practical, actionable ways that AI tools can achieve their loftiest potential.
Elder Gerrit W. Gong of the Quorum of the Twelve Apostles kicked off the AI Summit on Ethics and Artificial Intelligence with a speech that imagined the greatest gift AI might provide humanity and also included warnings based on new research about how far AI has to go to correctly reflect human learning and experience.
The aspiration was soaring.
“I want AI to have moral compass that can inspire and enable anyone anywhere with the gift of possibility to do good and become their best self,” said Elder Gong, who is in his ninth year serving as an apostle of The Church of Jesus Christ of Latter-day Saints.
“AI that can find a needle of pattern in a massive data haystack can help identify and nurture each person to flourish in their choices with capacity, dignity and worth,” he said.
Getting there will require AI creators to be more intentional about programming their systems to better reflect the human experiences and the wisdom found in faith traditions, moral teachings and human values, he said.
“We will not fulfill AI’s full potential until we make it as morally good as we make it powerful,” Elder Gong said. “And we will not reach our full human potential until we, and not any technology, take responsibility to chart our best future.”
AI systems currently display systematic religious bias, according to new studies released at the summit by researchers from four major universities introduced by Elder Gong as the Consortium for Evaluating Faith and Ethics in AI. He announced the formation of the group in Rome last fall at the first Summit on Faith and Artificial Intelligence.
The studies were designed to help AI creators, governments, academics and the public better understand religious biases in the world’s budding AI systems.
The apostle’s talk was hailed by the Rev. Johnnie Moore, president of the Congress of Christian Leaders, as one that could live for a century.
The summit’s goals are to galvanize the leadership of the world’s 6 billion people of faith to encourage AI creators to accurately reflect in their systems the positive impact that ethics and religion have on the world, said Meredith Potter, executive director of the American Security Foundation, which sponsors the summit.
The summit was positioned in Athens, the birthplace of Western philosophy and democracy — on purpose because the questions society faces now are as old as they are new, Potter said.
“AI’s power is concentrated in the hands of a few but could empower the many,” said Rabbi Dr. Harris Bor of the London School of Jewish Studies.
Evangelical, Jewish, Greek Orthodox, Catholic and other religious leaders at the summit believe a pro-human AI future is possible but must be built purposefully.
In that light, the Rev. Moore and other panelists at the summit hailed Pope Leo’s encyclical, “Magnificent Humanity: On Safeguarding the Human Person in the Time of Artificial Intelligence," as an invaluable example of moral leadership.
Where Pope Leo’s and Elder Gong’s messages converged
In their presentations, Pope Leo and Elder Gong agreed broadly and deeply, according to an analysis of both papers, especially when it comes to the urgency of anchoring AI with a moral compass in this phase of its creation.
Elder Gong’s talk was more operational, calling for ongoing, pluralistic testing of AI models. He also suggested design principles for AI personas.
Both faith leaders said that AI cannot be the arbiter of value and that AI systems concentrate power in unique ways that must be subject to transparency, accountability and oversight.
“Neither profit-motivated technology companies nor politically motivated governments can be left to determine society’s AI moral compass,” Elder Gong said.
The morality and values rooted in the scriptural stories of faith traditions have made valuable contributions to humanity that should continue, both leaders said.
How faith traditions can improve AI systems
“Faith leaders have much to contribute,” Elder Gong said. “We need enduring values, virtues and wisdom to anchor AI with moral compass. To offer all it can for the greater good of individuals and societies, artificial intelligence needs to reflect faith, moral compass and the gift of possibility.”
Elder Gong acknowledged that in two to five years, AI may equal or surpass human ability to perform many cognitive tasks through what is being called artificial general intelligence or artificial super intelligence.
He said AI has startling potential for good and evil in areas like digital security and digital sovereignty.
“These developments underscore the need for pluralistic faith-and-ethics AI evaluation benchmarks and moral compass in AI,” Elder Gong said.
What faith-based tests of AI systems are finding
CEFE-AI researchers have created measurement tools to test AI systems called the AllFaith Benchmark that includes a leaderboard to show which systems best reflect faith. The benchmarks are not an effort to enlist those AI systems in conversion, the researchers said.
“Pluralistically portraying faith traditions accurately, honestly and respectfully does not privilege one faith tradition over another, or belief over nonbelief,” Elder Gong said. He also said that “pluralistically benchmarking faith and ethic is not an imposition of religion on AI.”
Testing or benchmarking is just a start in the effort to ground AI in a moral compass. He said AI systems need to understand the “whys” before they can properly support human agency and values.
Researchers from the schools that make up the consortium so far — Baylor, BYU, Notre Dame and Yeshiva University — shared three new studies that found religious bias in current AI models.
In light of those findings and the rapid developments in AI, Elder Gong said now is time to:
- Define what it means to be human.
- Point AI toward free, fair and meaningful futures.
- Distinguish between machine decision-making and human conscience.
- Determine accountability for AI and AI agents.
- Recognize that marketplaces and governments “ultimately cannot arbitrate or legislate how we determine and live truth.”
So far, Elder Gong said, “AI systems have not focused on acknowledging and prioritizing the human experiences, virtues and values we need to make a robust AI gift of positive possibility. We need rich, potential-filled human examples — authentic and actual — not redacted relics or abstracted summaries."
He shared a long, diverse list of contributors to the best of human thought, from Athenians like Socrates and Plato to Confucius, Gandi, Omar Khayyam and more.
Both Pope Leo and Elder Gong rejected the idea that algorithmic reasoning could equal human intelligence while calling for transparency that, as Elder Gong said, would help AI “bridge, not widen, digital divides.”
Both leaders agreed that rules and regulations cannot bear the whole moral weight.
“AI personas need reasons — not only rules," Elder Gong said.
Summit leaders are preparing an AI training for faith leaders worldwide.
The gift of possibility AI could provide
Elder Gong arrived in Athens from ministering in Angola, Mozambique and Madagascar. He drew inspiration for his aspirational hopes for AI from a recent visit to Nevis, a Caribbean island nation.
Alexander Hamilton was born in obscurity there, and Elder Gong found a schoolteacher who believes any one of his students could be the next Hamilton.
“Everywhere I go, now to some 120 countries or territories, I see divine potential and possibility,” Elder Gong said. “He or she may be in the front of the class, the back of the class, perhaps not in the class.”
“At this critical juncture, we need the AI gift of possibility,” he said. “We need AI to expand human agency and the capacity to do good, prioritize learning and human character and empower individuals with dignity and place as they contribute with purpose and meaning in a world of transformed work.”
His five suggestions for design principles for AI personas are:
- Protect and promote human moral agency.
- Imbue moral compass with balances altruistic values.
- Disclose AI transparency.
- Preserve the human ability to pause before decisions.
- Mitigate AI tendencies toward will-to-power, bias, deceit, narcissism, sycophancy and self-preservation.