Artificial intelligence stands as one of the most pivotal advancements in modern technology. Its applications range from composing emails and tutoring students to offering companionship in times of need. A recent study in the Harvard Business Review forecasts that by 2025, generative AI will most commonly serve in roles related to therapy and social support. People are turning to AI for guidance on personal struggles, posing questions they might have historically directed toward mentors, advisors, or spiritual leaders: “How do I forgive betrayal?” or “How do I manage my anxiety?”
The answers they receive can often be disheartening. In the best cases, responses amount to vague advice like “consider mindfulness” or “connect with your values.” In more troubling instances, the guidance provided is not just generic but could potentially threaten lives. The unsettling reality is that AI has assumed the role of a significant spiritual advisor in America—yet it operates devoid of any genuine belief system.
My team at Gloo has put forth the Flourishing AI Christian (FAI-C) Benchmark, which evaluates how effectively leading AI models promote human flourishing within a Christian context. Our assessment examined seven essential dimensions: finances, character, happiness, relationships, meaning, faith, and health. Alarmingly, the Faith dimension received the weakest score, averaging just 48 out of 100 among the 20 AI models evaluated. Many of these models struggled to articulate key Christian principles such as grace, sin, forgiveness, and biblical authority, often substituting vague ideas for concrete guidance.
The implications of these findings are critical. AI has not emerged as a hostile entity toward Christianity; rather, it has been intentionally neutralized. Trained on a foundation of predominantly secular data and optimized to avoid offending anyone, these AI systems tend to deliver the most generic forms of spirituality. Therefore, their language may sound well-meaning, but it fundamentally lacks depth. This is a worry because AI does more than answer questions; it molds our belief systems. If the upcoming generation relies on AI for moral direction and receives little more than platitudes, the consequences extend beyond a loss of theological understanding. There is a risk of eroding the very framework for moral development.
For many Americans, faith plays a significant role—not merely as a lifestyle choice, but as a core pillar of meaning and purpose. When AI consistently overlooks this critical foundation, it is not acting neutrally; it is taking a definitive stance.
To alter this course, two fundamental changes must occur in the development of AI models. First, these models should be proficiently trained to engage with faith as seriously as they do with other subjects such as science, history, or literature. The intention is not to preach but to respectfully and accurately interact with the worldviews their users hold. Second, there must be robust benchmarks to evaluate this engagement. Without proper measures, accountability falters, and without accountability, improvement remains elusive.
The creation of the FAI-C Benchmark aims to shine a light on where current AI frameworks fail to grasp the needs of a diverse user base. Indeed, the stakes involved are substantial. When utilized correctly, AI has the potential to disseminate wisdom and foster community ties, ultimately promoting authentic human growth. Conversely, if mishandled, much like the unbridled challenges faced in the realm of social media, it risks accelerating moral decline. This would result in a culture that prioritizes comfort over conviction and sentiment over truth.
Strong moral structures are essential for a thriving society, and for billions globally, Christianity provides that framework. If AI cannot respect, recognize, and engage with this reality, it runs the risk of becoming a means of cultural diminishment rather than elevation. The aim here is not to transform AI into a preaching device; it is to prevent it from erasing crucial perspectives.
As AI technology evolves, it is crucial that it becomes more attuned to the nuances of the human experience. The central question is not whether AI will influence the next generation, but whether that influence will be constructive or detrimental.
"*" indicates required fields
