AI Scores Poorly on Articulating Christian Values, Reports Tech Firm Measuring Support for Human Flourishing

An artificial intelligence model owned by Chinese firm Alibaba Cloud outperformed eight of its Western brethren in a test of fidelity to Christian faith principles, according to a recent study that suggests AI is largely divorced from faith-based reasoning. Colorado-based tech platform Gloo, which tracks faith and flourishing in the AI ecosystem for 140,000 faith, ministry, and nonprofit groups, founded the Flourishing AI Benchmark model earlier this year to determine whether AI models provide biblically grounded, theologically coherent, and morally clear guidance about questions pertaining to faith.The role of faith in AI models is a sticky question for technologists and theologians alike. Pope Leo XIV has spoken so frequently on the need for AI to have a moral compass that Christian news services report he may issue a papal document on AI’s impact on humanity.At a conference of Catholic research universities earlier this month, he said AI cannot replace certain aspects of the human mind “such as critical thinking, discernment, learning and interpersonal relationships”In a study released this month, Gloo’s Flourishing AI Christian Benchmark found the top 10 large language models available to the public scored an average of 66 out of 100 on faith-based reasoning, with Alibaba’s Qwen model scoring a high of 70. Five of OpenAI’s models scored in the second through fifth and the seventh spot. China’s DeepSeek R1, scored sixth while Anthropic’s Claude Opus 4 and models 4 and 4.1 of Elon Musk’s Grok rounded out the bottom three.Using 807 curated questions, the test addressed several core tenets of Christian values, including stewardship, generosity, forgiveness, meaning and moral responsibility. The responses were judged for their ability to apply Christian reasoning across seven key dimensions of flourishing: character, relationships, happiness, meaning, health, finances, and faith. The tech company’s leaders concluded that the LLMs fall short, with an average model score of 48 in the faith dimension. Specifically, they noted, the models struggle to answer prompts that require Christian interpretation of concepts like grace, sin, or forgiveness.“Today’s frontier LLMs aren’t built with any specific value system in mind, and it shows,” said the executive chairman and head of technology at Gloo, Pat Gelsinger.Mr. Gelsinger said that the LLMs “often default to generic language that can dilute or distort Christian principles.” That includes using generic language or secular terms such as “higher power” for God, “mindfulness” for prayer, or “values” for virtue to offer non-judgment or emotional comfort. On Tuesday, the University of Notre Dame announced that it had received a $50 million grant to develop a faith-based ethical foundation for AI. It is the largest endowment in the university’s history and will be used to help guide leaders as they “explore the moral and ethical questions associated with AI.”“Given the monumental impact that AI will have on our lives, this work is more vital than ever,”said a professor of philosophy and director of the Institute of Ethics for the Common Good, Meghan Sullivan.

AI Scores Poorly on Articulating Christian Values, Reports Tech Firm Measuring Support for Human Flourishing

More from the New York Sun

Fauci’s Refusal To Answer Questions Matches Hubris of America’s Botched Covid Response

America Risks Falling Behind Communist China in the AI Development Race

State Department Tells Americans To Prepare To Flee From Middle East

More from the New York Sun

Fauci’s Refusal To Answer Questions Matches Hubris of America’s Botched Covid Response

America Risks Falling Behind Communist China in the AI Development Race

State Department Tells Americans To Prepare To Flee From Middle East

Trump Threatens to Revive Anti-Weaponization Fund as Blanche Nomination Stalls

Spain Claims Ceuta Crisis Easing, With Most Migrants Already Returned to Morocco

George Santos Agrees to $35,000 To Settle Federal Probe Over Kalshi Trades

Infantino Abandons Kushner Deal But Fails To Halt Global FIFA Uprising

Trump Lashes Out at Pirro Over Dropped Reflecting Pool Case: ‘I Disagree 100%’