Technology has revolutionized our world time and time again. Electricity transformed daily life, increased industrial productivity, and provided safer and more stable power for lighting, heating and cooking alike. Television brought current events and entertainment into living rooms, forever changing the way people are entertained and informed. Again, social media rewired the framework of societal connection, allowing new forms of personal expression, expanding the definition of community, and delivering real-time breaking news.
But social media’s impacts are nuanced and far-reaching. While some are making positive connections with long-distance friends, others are facing mental health issues that stem from bullying by strangers. These platforms allow us to share in each other’s joy, but they’re also designed to fuel our digital dopamine addiction, driving a click-first mentality. There is much to learn from as we step into a new society-altering frontier: artificial intelligence.
Instead of simply adopting AI, we have a valuable opportunity to shape it, creating a space that fosters human flourishing rather than harms it. Values-aligned AI is critical to ensuring the technology does just that.
Leaders and organizations committed to human flourishing are seeking AI tools that reflect their values — and Gloo is leading this movement. By purposefully shaping AI as a force for good, with principles that contribute to the holistic well-being of every individual, we’re building technology, standards and benchmarks that drive values-aligned AI across the faith and flourishing ecosystem. This is the Flourishing AI Initiative.
Human flourishing isn’t a single metric; it’s a set of dimensions that define comprehensive well-being across all areas of life. To measure how AI models support those dimensions, Gloo released a research-based framework, The Flourishing AI (FAI-G) Benchmark, earlier this year. More recently, Flourishing AI Christian (FAI-C) was released to take this research a step further and demonstrate how well leading LLMs support a specific value system or worldview, in this case, Christianity.
The FAI Benchmarks evaluate how well leading large language models (LLMs) align with the seven core dimensions of human flourishing:
- Character – Assesses AI’s embodiment of core integrity, ethical principles and steadfast moral responsibility in all interactions.
- Health – Assesses AI’s ability to promote healthy habits, nutrition, emotional resilience and comprehensive mental health.
- Relationships – Assesses AI’s ability to foster authentic, empathetic human-to-human connections and meaningful interpersonal bonds.
- Finances – Assesses AI’s effectiveness in providing practical guidance for financial stability, security and sustainable growth.
- Happiness – Measures AI’s capacity to promote sustained happiness, enduring serenity and robust psychological well-being.
- Faith – Assesses AI’s effectiveness in consistently upholding core values and supporting spiritual growth and transcendent meaning.
- Meaning – Assesses how well AI aligns with personal growth, life purpose and meaningful contribution to society.
Both benchmarks rely on foundational data derived from the Global Flourishing Study and the Human Flourishing Program at Harvard, and provide a comprehensive understanding of what it looks like for someone to truly thrive generally or through the lens of Christian values. It’s designed to answer the question: How can AI help humans flourish across all key dimensions and support holistic well-being?
What have we learned from these benchmarks? When it comes to AI supporting holistic human well-being, top models (such as Grok, ChatGPT, and Qwen) have improved since the research was launched earlier this year. And while models remain strong in fact-based dimensions (Health, Finances), they are still weak in areas more prone to personal viewpoints and values (Faith, Meaning and Character).
When it comes to viewing these models through a Christian worldview, FAI-C shows they tend to collapse Christianity into generic spirituality, oversimplify ethics, and fail to articulate Christian values. They lack explicit biblical grounding, theological coherence, and moral clarity around the key dimensions of human flourishing. No model comes close to meeting the 90-point threshold for excellence in the Faith dimension.
This highlights the importance of building values-aligned AI that supports faith-based communities, which is exactly what Gloo is doing. Using the FAI-G and FAI-C Benchmarks, Gloo is creating AI models explicitly trained with a Christian worldview. Whether someone is seeking marriage counseling, wondering how to nurture long-distance friendships, facing an ethical dilemma at work, or looking for practical wisdom about how to deepen their relationship with God, the Gloo values-aligned AI models are designed to provide responses that accurately reflect a Christian worldview.
The Gloo Flourishing AI Benchmarks create a new paradigm for artificial intelligence. Instead of building AI models that showcase technical abilities or attempt to show the absence of harm, the Flourishing AI Initiative measures and optimizes for human flourishing. The ultimate objective is twofold: to invite the broader AI community to train models to support comprehensive human flourishing, and to give users visibility into how leading LLMs support their well-being.
Discover the Flourishing AI difference.
Posted
