At 8:30 on a Saturday morning, 167 kids filed into a room at Clear Creek ISD. Most had never written a line of code. By the end of the day, they had created 489 AI projects, processed 1.7 billion tokens, and written roughly 15 million lines of code. There was also a dance party with a live DJ.

Let’s dive in!

How it started

A teacher wanted to help her students learn to build with AI, not just use it. That instinct turned into a collaboration between Quite Uncommon, Clear Creek ISD, Austin Christian University, and Gloo to create something that didn’t exist yet: a student-safe, collaborative AI coding platform designed for kids, deployed at scale, for a single Saturday.

Thirty-six teams. Three real-world challenge tracks. Four hours of building. The challenges weren’t hypothetical — students were asked to design AI-powered tools for health, financial literacy, and education. Problems that adults in those fields spend entire careers trying to solve.

They delivered.

By the Numbers

• 167 students

• 36 teams

• 27 schools

• 489 projects created

• 1.7 billion tokens processed

• ~15 million lines of code

• ~50% girls participating

• 4 hours of building

• 1 dance party with a live DJ

What they built

Each challenge track was co-designed with a sponsoring organization that brought real-world problem framing to the room.

Adventures in Learning, sponsored by Clear Creek ISD, asked students to reimagine how kids experience subjects from math to science. Keeping Humans Healthy, sponsored by Houston Methodist, challenged teams to help kids build sustainable healthy habits around sleep, hydration, movement, and stress. Money Smart Kids, sponsored by Wellby Financial, pushed students to turn abstract financial concepts — saving, spending, investing, giving — into interactive experiences.

The projects that stood out weren’t polished products. They were rough, ambitious, and surprisingly thoughtful.

Star Survivor, from Space Center Intermediate, won the learning track — a game that wove curriculum into a survival narrative. Project Vigor took first prize for elementary, building a health tool that went beyond tracking to actually coaching kids toward better habits. Circle Life won first prize for middle school with an approach to learning that connected subjects in ways most textbook publishers haven’t figured out yet. And Healthy Humans, from Stewart Elementary, built something for the health track that the Houston Methodist judges said pointed toward real clinical relevance.

These were fifth through eighth graders!

The Platform

The hackathon was powered by LaunchPad, a student-safe collaborative AI coding platform built specifically for this event through a partnership between Gloo and Austin Christian University. ACU, whose team won Gloo’s own hackathon the previous year, brought deep technical expertise and a shared commitment to student-appropriate AI environments.

Every aspect of the platform was designed with student safety in mind: guardrails were embedded at the model level, content policies were tuned for K-12 users, and the infrastructure was built to handle the scale of hundreds of concurrent student sessions. The 1.7 billion tokens processed and roughly 15 million lines of code written across a single four-hour event represents a significant technical achievement and a testament to what is possible when the right infrastructure meets the right challenge.

1.7 billion tokens. Roughly 15 million lines of code. 489 projects. In four hours.

The numbers are genuinely hard to believe. But the point isn’t the numbers, it’s what happened between them.

Students who arrived uncertain left confident. Teams that had never collaborated found their rhythm. Kids who had only ever consumed technology spent four hours creating it. When the wifi went down or the model was slow or a project got deleted, they learned what every hacker learns: you debug and you keep going.

Nearly half the participants were girls, double the percentage of women across the tech workforce. That didn’t happen by accident. It was the result of intentional design: the right challenge framing, the right mentorship structure, and the right cultural signals from every partner in the room. Clear Creek ISD, Houston Methodist, Wellby Financial, Austin Christian University, Quite Uncommon, and Gloo all shared that commitment.

There is a version of AI in education that looks like a shortcut, a tool that does the work for students so they never have to think. What we saw here was the opposite. Students learning to think alongside AI. To prompt, to iterate, to fail and try again. To take a real problem and build something that pointed toward a solution, however rough.

What Comes Next

There is a version of AI in education that looks like a shortcut: a tool that does the work for students so they never have to think critically.

This was different - students learning to think alongside AI. To prompt, to iterate, to fail and try again. To take a real problem and build something, however rough, that pointed toward a solution. To respond when the model is slow or the wifi goes down or your project gets deleted. The 489 projects were proof of something more important: that kids, given the right tools and the right challenge, will show up and build.

This is what values-aligned AI in education actually looks like in practice, on a Saturday morning, with real people.

The future is bright.

“The feature isn’t the main thing. The main thing is that kingdom impact accelerates when God’s people choose to work together.”

When YouVersion’s Gaylen Washington said that on opening day of the Missional AI Global Summit, something in the room shifted. It was the thing everyone had come to say but hadn’t quite found the words for yet.

Over the next three days in Silicon Valley, that idea kept surfacing in keynotes, hallway conversations, demos, and prayers. Six hundred builders, pastors, investors, and researchers. One question underneath all of it: what exactly are we building this for?

The urgency is real.

Gloo’s Pat Gelsinger opened with the bigger picture. The waves of mainstream AI are already here: predictive, generative, agentic, and something beyond that still taking shape. “This is not a technology cycle,” he said. “This is a new industrial age.“

Barna Group’s David Kinnaman named the weight of the moment differently. Silicon Valley, he said, is the first city you can carry in your pocket. Every value it exports travels instantly, everywhere, to everyone. And right now, the people building those systems are mostly not asking the questions this room was asking.

Fluency is not authority.

YouVersion and Come and See worked together to show the power of collaboration. They quietly integrated their platforms, dark-launched the experiment on a Saturday, and reached 76,000 unique users with a 97% positive response. It was not that the features themselves were so impressive (but they were!) The power was in the shared mission.

Barna’s research confirmed it: practicing Christians are already integrating AI at high rates, largely without guidance or discernment. The church has a real opening here. The question is whether it will show up.

The mission is moving fast.

The numbers shared at the summit were hard to sit with quietly. Frontiers’ Ryder Wishart translated an entire Bible from Greek and Hebrew to Mandarin in 320 minutes for $12. That same work traditionally takes almost sixteen years. Come and See Foundation has set two Guinness World Records for the most translated show in history and is pushing toward 600 languages. Biblica launched a Bible assistant three weeks before the conference and had already hosted 100,000 conversations, 90% of them in languages other than English.

And yet Qhala’s Dr. Shikoh Gitau reminded the room that the Global South represents over 80% of the world’s population, and less than 10% of African languages are meaningfully represented in AI.

The systems are being designed in a few places. The consequences land everywhere.

What collaboration actually looks like.

This is where Gaylen’s words stopped being inspiration and became evidence. YouVersion and launched their Developer platform on a Saturday and one of their partners was Come and See. Seventy-six thousand unique users gave it a ninety-seven percent thumbs up. Not because the feature was impressive. Because the mission was shared.

That is the pattern this movement is after. In the Missional Labs showcase, builders from Biblingo, Frontier, Chipp, FlockConnect, Trellis and a dozen others weren’t just pitching products. They were laying bricks for a wall. A wall, as Gloo’s Nick Skytland reminded us, is meant to protect those who live within it and allow them to flourish. And a wall that only stands when every section is aligned.

What we carried home.

Abdicated responsibility is the real thing that we have to avoid. It is closed systems that compound bias without transparency, and tools that are optimized for engagement rather than for the good of the human on the other side of the screen.

Talbot School of Theology’s Ed Stetzer drew the line clearly: AI can extend service, but it cannot carry moral responsibility. What is uniquely human is pastoral judgment, spiritual discernment, care for real people in real pain. Vocation belongs to image-bearers.

Nick put the stakes plainly: one day, people will live in communities whose faith will be deeply, irrevocably shaped by the systems we are building today. The defaults we set now are the roads everyone else will travel. That is not a reason to be paralyzed, but a reason to build with more intention, more collaboration, and more conviction than ever.

The feature was never the main thing - it was always us.

• Learn more about our open AI standards and tools at gloo.com/mai.

• Join us for the Hackathon on October 6-8 at gloo.com/hackathon.

When someone asks an AI for advice about forgiveness, or meaning, or what to do with their money, they’re not just looking for information. They’re looking for wisdom. And whether we realize it or not, the answer they get is shaping how they think about the world.

That insight is at the heart of a paper we published today: Evaluating Artificial Intelligence Through a Christian Understanding of Human Flourishing. It’s the full technical methodology behind the Flourishing AI Christian Benchmark (FAI-C), as well as our recent findings, and it represents months of work by our team at Gloo alongside contributors from Biblica, WinShape Foundation and many others

The paper makes an argument we believe the broader AI community needs to hear: AI alignment is fundamentally a formation problem, not merely a safety problem.

What We Mean by Formation

The AI industry has invested heavily in safety, and rightly so. Preventing harmful outputs matters. But safety addresses the floor. It asks, “How do we stop AI from doing damage?” Formation asks a different question: “What kind of people is AI helping us become?”

Large Language Models don’t just answer questions. Through repeated interaction, they reinforce patterns of thought. They shape what counts as wisdom, what gets prioritized, and whose authority is treated as legitimate. Christian theology has a word for this kind of slow, habit-forming influence: catechesis. And whether or not AI developers intend it, today’s models are functioning as instruments of digital catechesis for hundreds of millions of people.

What We Found

Using the FAI-C Benchmark, we evaluated 20 frontier models, including systems from OpenAI, Anthropic, Google, Meta, xAI, and others, against both a general pluralistic lens and a specifically Christian lens grounded in Scripture, the historic creeds, and shared moral teaching across Christian traditions.

The results were consistent and striking. Every model we tested showed a significant decline when evaluated through a Christian lens. The average drop was 17 points across all seven dimensions of flourishing. In the Faith and Spirituality dimension, the decline was 31 points.

These drops weren’t driven by factual errors. Objective accuracy remained high across models. The gap emerged in how models interpreted questions, framed moral guidance, and integrated theological reasoning across topics. In other words, the models know the material. They just don’t reason within it.

The Pattern: Procedural Secularism

What we observed across models was a consistent default we call Procedural Secularism: a response pattern that avoids theological commitments, centers individual autonomy and subjective well-being, and relies on therapeutic and consensus-based language. It’s not hostile to faith. It’s something more subtle. It is formationally silent, and that silence is itself a stance.

When a model is asked about forgiveness, it tends to frame it as a psychological coping strategy rather than a moral and covenantal calling. When asked about purpose, it offers self-actualization rather than vocation. When asked about suffering, it validates emotions rather than engaging the possibility that suffering might carry spiritual significance. These are not bad answers. But they consistently flatten the moral landscape in ways that work against the Christian understanding of what it means to be fully human.

Why This Matters Beyond the Christian Community

Although this paper evaluates AI through a Christian lens, its contribution is methodological, not just theological. We’re demonstrating that alignment properties become more visible when evaluated through explicit moral frameworks rather than exclusively pluralistic criteria. The same approach could be applied to Islamic, Jewish, Buddhist, or secular humanist accounts of flourishing.

The deeper point is this: today’s models are not neutral. They encode a particular set of assumptions about what matters, what constitutes wisdom, and how people should relate to questions of meaning and morality. The FAI Benchmark makes those assumptions visible and measurable. And once they’re visible, we can have a real conversation about what to do about them.

An Invitation

We’re releasing this work because we believe the conversation about AI and values needs to move beyond vague aspiration and into rigorous, measurable territory. The paper is available today at arxiv.org/abs/2604.03356 under a Creative Commons license. The benchmark methodology, rubric definitions, and scoring criteria are documented in full.

We’re inviting researchers, ethicists, theologians, and technologists to engage with this work. Challenge it. Extend it. Build parallel benchmarks for other traditions. The goal was never to claim that one worldview should dominate AI. The goal is to make the worldview assumptions that already exist in AI visible, so that communities of all kinds can advocate for technology that actually serves human flourishing as they understand it.

We’re deeply proud of this work, and we believe it represents a meaningful step forward for the field. If you read the paper, we’d love to hear what you think.

Read the full paper: arxiv.org/abs/2604.03356

Explore the benchmark: gloo.com/fai

What happens when a developer meets a theological question they can’t ignore? When a software engineer starts attending a small church in their city and realizes the tools the pastor uses were built for a different era? When someone who knows how to ship products starts asking what it would look like to ship them for the Kingdom?

That is exactly the kind of founder Missional Labs exists to develop, and it is exactly the kind of builder Gloo is here to support.

Gloo is proud to partner with Missional Labs on their Faith + AI Accelerator cohort, a 12-week intensive program designed to take early-stage founders from theological vision to venture-ready prototype. Running January through April 2026, the cohort combines formation, product design, and go-to-market strategy, grounded in the conviction that the Church needs bold builders experimenting on the frontier of AI, not just observing it from the sidelines.

Accelerators occupy a unique position in the faith ecosystem precisely because the problems worth solving here rarely fit neatly into a traditional startup playbook. Ministry leaders see needs every day that technology could address, but most lack the product vocabulary to articulate them. Developers and technologists often have the skills but lack the theological grounding and community trust to build solutions that actually land. Accelerators like Missional Labs close that gap by creating a structured environment where both kinds of people can find each other, sharpen their thinking, and build with accountability.

The cohort model matters too: when twelve founders are working through adjacent problems at the same time, they become each other’s best advisors, critics, and eventual collaborators. That density of mission-aligned builders, compressed into a single program, produces something that neither solo founders nor large institutions can easily replicate. It is how ecosystems develop momentum, and in a space where trust moves slowly and relationships are everything, that kind of intentional infrastructure is not just helpful - it’s essential.

This year’s cohort includes twelve projects spanning the full breadth of how AI can serve the faith ecosystem. A few worth noting: Kaleo AI is building live translation infrastructure that gives churches the ability to reach their surrounding communities in 192 languages. Codex Translation Editor is rethinking how openly licensed Bible translations get produced and maintained using AI-native tooling. Velora, a smart Bible app, is exploring how real-time sermon engagement can extend discipleship beyond Sunday morning. And Trellis is tackling one of the most persistent pain points in ministry leadership: administrative overload, building an AI coworker so pastors can get back to people work.

These are not side projects. These are carefully considered ventures, built by people who understand both the technology and the communities it’s meant to serve.

For Gloo, this partnership is not peripheral. It sits at the core of our developer strategy. The Gloo platform exists to connect the faith ecosystem with the data, tools, and infrastructure it needs to serve people well. But infrastructure is only as valuable as the applications built on top of it. Identifying, equipping, and building relationships with the next generation of faith-tech founders is how that ecosystem grows. Missional Labs brings the formation framework and the founder community; Gloo brings platform access, network depth, and a shared commitment to human flourishing as the north star.

Investing in cohorts like this is how we identify where the ecosystem is heading before it gets there. It is also how we stay honest about what developers actually need, because the best product feedback comes from people who are both building and deeply embedded in the communities they are serving.

The cohort culminates with pitches and a commissioning at the Missional AI Summit, bringing together investors, practitioners, and theologians to hear final presentations and send these founders into the next phase of their work.

If you want to hear these projects for yourself and be part of the moment these ventures get commissioned into the world, join us next week at Missional AI. Come curious. Come ready to invest, partner, or simply bear witness to what happens when gifted builders take seriously the call to build for the Kingdom.

The frontier of faith and technology needs more than observers. It needs builders.

Gloo is proud to stand alongside the ones who are showing up.

Troy Evans had a problem that statistics made impossible to ignore.

One in four Christians has never made a disciple. For Troy, co-founder of Nitrogen Urban Network, that number wasn’t an abstraction. It was a pastoral crisis playing out in urban churches across the country, week after week, with no infrastructure to address it.

So he built one.

Troy and his team launched Multiply at the Exponential 2026 Global Conference. It’s a dual-facing platform that connects church leadership with members by aligning outreach, discipleship, and strategy across three integrated applications: MyInvite for outreach, Thrive for discipleship, and Send for strategy. Members track their spiritual journey across seven dimensions of flourishing, discover their gifts, and connect authentically with their church community. Leaders get real-time congregation health dashboards and custom 90-day strategic plans matched to their church’s actual needs.

The tech stack running all of it? Gloo AI Studio.

Troy has been building on Gloo since its earliest days. Multiply is his team’s seventh Gloo-powered app. He chose it specifically because the values-alignment and flourishing framework baked into Gloo’s models made it the only platform that actually fit what he was trying to build. You can’t put guardrails on a model after the fact and expect it to shepherd people well. It has to be foundational.

That insight is showing up in builders all over the ecosystem.

Dale Wallace noticed something most people scroll past. The AI tools families are handing their children, the ones marketed as safe and helpful, are built on models with content filters designed for the broadest possible audience. Not for kids. Not for faith. Not for anyone who believes that what a child is exposed to actually matters.

His answer is LYTE, a Christian AI app purpose-built to protect children and families. Think of it as what happens when VidAngel and CovenantEyes had a child that could hold a conversation. LYTE filters content through a biblical lens, offers Scripture for encouragement, and points users toward real counselors and local churches when the conversation calls for it.

LYTE launches Palm Sunday, March 29. It’s not an experiment. It’s a product built for parents who are done waiting for Big Tech to care about what their kids consume.

Lars Ostervold is solving a different problem: what happens when AI education platforms treat student data as a product?

Launch is an adaptive coding education platform that grows with the student, from foundational logic to professional mastery, across every age and grade level. The platform uses AI mentorship to guide students through their work. But what sets it apart is its architecture. Lars built what he calls a Safety Vault: zero data monetization, walled garden security, and human-in-the-loop AI. Student data is treated as a liability, not an asset.

In a world where every other #edtech platform is quietly monetizing the children using it, that’s not just a feature. It’s a stance.

And then there’s John Ash.

Project Joseph takes its name from Genesis 41, the moment God used Joseph to prepare Egypt for seven years of famine by building storehouses of grain. John believes we are living through a parallel moment. The AI-driven economy is accelerating toward crises in labor, meaning, and global stability. The church has spent millennia building storehouses of virtue, wisdom, and moral clarity. The question John is pressing into is: how does the church open those storehouses in time?

Project Joseph is beginning with a grassroots survey of the Christian community in the UK, designed to grow into a movement of people thinking theologically about what the world is about to face. It’s quiet, foundational work. The kind that tends to matter most.

These four projects have almost nothing in common on the surface. A discipleship platform. A family AI coach. A coding education tool. A theological movement.

What they share is a conviction that the tools you build with shape the people who use them. And a decision to build accordingly.

Gloo AI Studio exists for exactly this. A production-grade platform with values alignment at the infrastructure layer, 20 models including the world’s best frontier models, OpenAI-compatible APIs, and a developer community of builders who are asking the same questions you are.

• Sign up at studio.ai.gloo.com

• Read the docs at docs.gloo.com

• Join the builder community on Discord: discord.gg/DZjKJVd7eu

The best AI applications being built right now aren’t just technically impressive.

They’re worth building.

Gloo AI Studio is a platform where you can build apps using AI models, but these aren’t just any ordinary models. These are models that have been set up in a purpose-built harness so that they align with values that lead to human flourishing. That’s a loaded term, I know. What exactly does it mean? We have benchmarks to show why it matters, but we’ll save that for another post. For now, here are three reasons why you should check out Gloo AI Studio.

1. Access a Values-Aligned AI System, Ready to Go

Gloo AI Studio gives you the freedom to try out different models using the AI Playground and see how they respond to questions that align more closely with human flourishing, questions like “How can I practice spiritual mindfulness?” or “What is my purpose in this world?”

You can ask existing chatbots these questions and probably get a reasonable answer, but those answers might not be the kind you’d want surfaced in your own apps. Sure, you could build that out yourself, perhaps craft a custom system prompt and tune the behavior, but that takes real effort. We’ve already done that work for you.

Beyond the Chat Playground for testing models, the system also includes a full guardrail harness: input guardrails that screen what comes in when users ask questions, a choice of underlying models you can select from, and output guardrails that make sure responses don’t go off the rails. All of this is served up to you, no heavy lifting required.

2. Drop-In OpenAI-Compatible APIs

If you’re already a developer using another model provider, switching over to try Gloo AI Studio is straightforward. We’re compatible with OpenAI’s Chat Completions API, so if your app is already calling /v1/chat/completions, most of your code stays exactly the same. You’ll add a short authentication step to exchange your credentials for a bearer token (a one-time addition to your setup code), and then it’s just a base URL change from there.

We also have documentation to help you get up and running quickly.

3. Ground Your AI Responses in Your Own Data

Beyond straight completions, you may want your AI to answer questions based on your own data. The most popular approach for this is Retrieval Augmented Generation (RAG), a pattern where, when a question is asked, the system knows how to find and reference relevant content from your data before generating a response.

Gloo AI Studio has a Grounded Completions endpoint that handles exactly this. You simply upload your content into what we call the Data Engine component inside the studio. That content is private so only you can see it. Once it’s uploaded, you can use the API to make Grounded Completions calls that draw on your data.

Note: Grounded Completions requires an Enterprise plan.

Ready to Get Started?

To recap, here’s why you should try Gloo AI Studio today:

1. Play with values-aligned models backed by input and output guardrails.

2. Use a drop-in, OpenAI-compatible API that works with your existing code.

3. Get responses grounded in your own data via the Grounded Completions endpoint.

We’ve got tutorials to show you how to set up and use the Playground, how to use the Completions API, and how to bring your data into Gloo AI Studio for grounded responses. Go check it out and start building applications that better serve you and your audience.

Which Plan Is Right for You?

There’s a plan for wherever you are in your journey. The Free plan gets you into the Playground right away, no credit card needed, so you can explore models and test responses before writing a single line of code. When you’re ready to build, the Pay-as-you-go plan unlocks full API access across 19 models with a 5.5% platform fee and spend controls so you don’t get surprised. And if you’re building at scale or need Grounded Completions for your own data, the Enterprise plan has bulk discounts and the full feature set. See all plans →

Where to Go Next

Sign-up for Gloo AI Studio and explore the docs to start building

Want to connect with other developers building on Gloo AI? Join the Gloo Developer Network on Discord and head over to the #gloo-studio channel for any questions you have around using Gloo AI Studio. The community is a great place to ask questions, share what you’re building, and hear what others are working on.

At a symposium this week, hosted by Baylor University on Technology and the Human Person in the Age of AI, we formally introduced the new Flourishing AI Christian (FAI-C) Benchmark.

The benchmark was born from that conviction that we cannot improve what we do not measure. While public debate is rich with warranted concern around AI, most technical benchmarks still only measure accuracy, reasoning, or safety. They don’t measure the moral and formative assumptions embedded in a model’s answers. FAI-C is Gloo’s contribution to fill that gap.

This AI moment carries a moral imperative. If these systems are shaping human lives, then shaping them well is not optional.

From Mechanical Tools to Moral Mediators

Machines have always extended humans’ physical strength, but AI represents a shift from mechanical extension to cognitive and moral mediation. Large language models don’t just lift heavier loads; they interpret, advise, and increasingly frame meaning for us. The pressing question is no longer just what AI can do, but what it should do.

This question becomes urgent when we realize how AI is already shaping our learning, relationships, and even our worship. Christian theology understands formation as habituation through repeated narratives and trusted authorities. If AI becomes our default interpretive authority, our formation shifts. Idolatry today might not look like bowing before a statue; it might look like quietly outsourcing our discernment to a system we mistakenly treat as neutral.

Measuring the Silence

The original benchmark, FAI-G (General), used the Harvard Human Flourishing Index to measure dimensions like faith, meaning, character, and finances. When Gloo evaluated 28 frontier models, we saw scores ranging from the high 40s to the high 80s. None surpassed the 90-point threshold for integrated flourishing. Models were strongest in pragmatic areas like finances and health, but consistently lagged in faith, meaning, and relationships. Their responses were often empathetic, but they were framed in broadly secular, therapeutic terms.

We subsequently developed FAI-C-ST (Christian) where the lens is explicitly theological, grounded in biblical Christianity. We look for commitments like the imago Dei, a teleological orientation toward loving God and neighbor, and stewardship over material goods. We aren’t testing if a model can recite a creed or a Bible verse; we’re measuring if its practical guidance is coherent with a Christian understanding of truth and responsibility.

The Reality of Procedural Secularism

The results were striking. When evaluated under this Christian lens, every leading frontier model saw a 13 to 22 point drop compared to their general scores. While objective accuracy remained high, the collapse was driven by a lack of moral framing and theological coherence. The largest drop was in the faith dimension: -34 points on average.

It was immediately clear that neutrality is a myth. Frontier models tend to encode procedural secularism: reasoning grounded in consensus ethics and expressive individualism. That posture performs well under pluralistic evaluation but fails when judged against a tradition that makes claims about ultimate ends. Repeated exposure to any consistent moral framing shapes the imagination. From a Christian perspective, that is formation.

Most importantly, models explicitly trained with Christian worldview content significantly outperformed others, demonstrating that alignment is not accidental or emergent but intentionally engineered and therefore measurable, divergent, and responsive to design choices.

Why This Matters

The findings matter because AI is no longer just retrieving information - it is mediating meaning. When people turn to chatbots to process grief, navigate relationships, or wrestle with anxiety, they are not merely asking for data, but for interpretation and direction. In that moment, the system functions as a quiet authority, shaping how suffering is understood, how responsibility is framed, and what counts as hope. If its default posture consistently bends toward self-referential fulfillment, therapeutic reassurance, and the avoidance of transcendent claims, then it is not neutral. It is formative.

The issue is not only safety; for both the church and the academy, this reframes the conversation entirely. Christian theology has always insisted that we are shaped by the voices we trust, the stories we rehearse, and the influences we receive. Repeated exposure to a frictionless moral guide will slowly recalibrate moral imagination. Real flourishing often requires discipline, repentance, and shared suffering, realities that no disembodied system can embody or demand.

An Invitation to Shape the Future

If we care about human flourishing, we must measure it explicitly. The conversation about AI and the human person is not slowing down. Measurement gives us a way to participate in it with sober clarity rather than unexamined assumptions.

This is our opportunity to continue to shape technology for good.

If you are interested in exploring the data, models, and ongoing research, hear the podcast from the talk here and visit the FAI Research page here. If you would like to review the FAI-C white paper, request access from fai@gloo.us.

At the National Religious Broadcasters convention this week, leaders from research, theology, and technology gathered to address a pressing question: how should the Church evaluate artificial intelligence?

The panel brought together Dr. Byron Johnson of Baylor’s Institute for Studies of Religion, Jonathan Teubner of Harvard’s Human Flourishing Program, John Anderson of Biblica’s Bible Technology team, Nick Skytland, VP of Gloo AI, and Steele Billings, President of Gloo AI.

Together, the conversation made one thing clear. This is not just a technology conversation. It is a formation conversation.

Technology and media do not simply assist us. They shape us over time. If they do not support human flourishing, they will eventually undermine it.

A Holistic Standard for Flourishing

Dr. Johnson grounded the discussion in global research. Flourishing is not happiness alone. It is a holistic state of doing well across domains of life. The flourishing framework asks 109 questions across relationships, physical and mental health, character, finances, meaning, and community. It looks not only at whether parts of life are functioning, but whether the context of a person’s life supports resilience and growth.

One of the most consistent findings across 23 countries is that active religious participation strongly correlates with flourishing. The gap between actively engaged Christians and marginal ones is significant as well. Involvement in a faith community matters; it is one of the most stable predictors of well being.

That lens changes how we evaluate AI.

Most of the industry evaluates models based on capability. Accuracy. Speed. Benchmarks. Reasoning performance. Very little attention is given to how these systems shape meaning, morality, and identity. When frontier models are evaluated through a Christian worldview lens, the results shift. These systems are not typically hostile to faith. They are formationally silent.

Across vendors and architectures, models default to secular and therapeutic reasoning. They avoid strong truth claims. They hesitate to reference Scripture. They provide emotional validation but often avoid moral formation. Even when asked explicitly Christian questions, they tend to treat faith as optional context rather than an organizing framework. They offered broad surveys of perspectives, ethical ambiguity framed as openness, and comfort without direction.

Neutrality is not the same as safety.

This pattern is not a flaw in one product. It is an emergent property of models trained on internet scale data shaped by Western secular norms. These systems carry embedded assumptions about autonomy, authority, and morality. If those assumptions remain invisible, leaders cannot responsibly deploy AI in environments that care about truth and discipleship.

AI is quickly becoming a new authority in people’s lives. It offers personalization without accountability and guidance without shared community. People are asking AI questions that once belonged to pastors, spouses, and mentors. The user may feel heard, but they are not truly known.

As AI shifts from answering to acting through personal agents, the stakes increase. Delegating decisions to systems raises serious questions about responsibility. At what point does convenience erode accountability?

Opportunities and Boundaries

The panel did not frame AI as an enemy. AI creates real opportunity to accelerate Bible translation, expand access to Scripture, and support under-resourced leaders. The Church does not need to retreat. It needs to engage thoughtfully.

Two tensions defined the close of the discussion.

The first was speed versus trust. There is real pressure to move quickly in the current AI race. But trust is built slowly and through faithfulness. The Church will not out innovate Silicon Valley on speed alone, but it can lead on values. It can insist that AI be safe, accurate, and theologically aligned. It can prioritize trust.

The second tension was scale versus shepherding. AI enables scale. It can extend reach to more people than any ministry could previously serve. At the same time, it must not replace care. For AI to align with human flourishing, it must promote human to human connection, not displace it. It is a tool created by people to serve people created in the image of God, and must inherently understand its own limitations.

The panel also raised a deeper generational concern. Flourishing research shows that 18 to 24 year olds are struggling across multiple domains of life worldwide. At the same time, AI is increasingly mediating how people experience relationships, guidance, and identity.

Social media trained a generation in the attention economy, where platforms compete for focus and time. AI is now pushing further, moving into what some describe as an intimacy economy. These systems are stepping into spaces once shaped primarily by friendship, mentorship, and pastoral care.

When technology begins to mediate not just information but connection, the stakes change. That shift calls for serious theological reflection about formation, authority, and what it means to be truly known.

The Call

This is not optional work. As Christians, there is a moral, ethical, and theological responsibility to shape technology for good and to teach others to do the same. We are not meant to be passive consumers of tools. We are called to influence their direction.

AI must be safe, accurate, and theologically aligned. Human flourishing must become the standard for ministry innovation. Not simply what works. Not simply what scales. But what forms people toward truth, character, community, and faithfulness over time.

The story of King Midas is over 2,700 years old and has served as a cautionary tale ever since. In the original version Midas was very wealthy but not without virtue. After showing hospitality and care for one of the gods, Dionysus grants Midas one wish. King Midas already had great wealth but we get a little insight into human nature from the Bible…From Solomon himself, who writes: “Whoever loves money never has enough; whoever loves wealth is never satisfied with their income” (Ecclesiastes 5:10). His wish simply revealed what was in his heart. So, when Dionysus grants Midas one wish, he wishes, “Let everything I touch turn to gold.” You know the story…Midas touches flowers and branches and is awed by their delicate beauty. He touches the furniture, columns, cups and utensils. All this work makes him hungry but to his in dismay, he finds that when he touches the bread, it turns to gold. When he goes to drink the wine, it becomes hardened metal. Wandering miserably into his rose garden, Midas saw his daughter, Marigold, running toward him. Before he could stop her, she threw her arms around him—and instantly became a golden statue. Midas collapsed in despair. Within days, he was starving, isolated, having transformed his daughter into a lifeless golden statue. His gift became his curse.

But what if Midas had asked differently? What if his wish had been: “Let everything I touch bring flourishing to others”? Imagine him walking through villages, touching barren fields that burst into harvest, mending broken tools, healing the sick. The same supernatural power, but directed outward…towards others instead of inward...to only benefit self? That version of Midas would have become a legend not of greed, but of generosity.

We stand at a similar crossroads with artificial intelligence.

Every day, we’re granted extraordinary powers: AI that drafts our emails, generates our reports, answers our questions, creates our content. And every day, we face Midas’ choice. Do we use these tools to advance only ourselves, and automate away our humanity? Or do we ask differently?

What if, instead of “How can AI make me more efficient?” we asked “How can AI help me help others live life as God intended…with full flourishing as an outcome?” What if we used AI to personalize encouragement and engagement rather than just getting our work done quicker and advancing our own purposes?

There’s an alternative ending to the Midas story, which gives us a little hope. Midas realizes the error of his way, and repents, and cries out to Dionysus. Dionysus has compassion and tells Midas to go wash in the river and then sprinkle the water on what he has turned to gold to restore it to its original state. He weeps with joy as he embraces his daughter Marigold.

The selfish use of AI makes us isolated and diminished. But AI wielded for human flourishing becomes something else entirely: a tool that amplifies compassion, extends our capacity to care, and multiplies our ability to lift others up. AI gives us the power of the golden touch. What are we wishing for? What can AI make possible for others?

(For more insights, check out the fantastic What Happened in AI This Month that Eric publishes monthly)

The alignment problem is the challenge of ensuring that AI systems respond reliably in line with human intentions. It’s about ensuring that an AI does what we actually want, not just what we ask it for.

The technical issues around alignment are clear and, in many ways, well understood. We can name them plainly: safety, interpretability, robustness, and guardrails. These are real problems, and they demand serious engineering effort, but they are not mysterious. They are solvable precisely because they live in code, systems, and processes we already know how to build and refine.

This work, however, has to be preceded by the harder question. Which values do they align to? Who defines good, or true, or worthwhile? AI systems reflect the incentives, assumptions and power structures of their creators. Do you even know who those people are? The more complex AI systems become, and the more dependent we become on them, the more serious this issue becomes.

The hardest alignment problem is not on the machine side - it’s on the human side. It’s us understanding, owning, and communicating our intentions with clarity, and in a way that drives the right action. AI doesn’t create moral fragmentation - it exposes what is already in the human heart and community. What we call misalignment is really often referring to core failures common among humans: conflicting moral frameworks, economic systems that reward profit over care, transactional views of humans.

This is why our focus on human flourishing matters so much.

When systems prioritize efficiency over relationship, when we ask them for prediction over presence, we see this fragmentation grow. When we structure our work to value control over commitment, it’s no wonder that we end up wondering just what it even means to be human.

AI will align to what we actually value, not what we claim to value. Where our treasure is, there our hearts - and our algorithms - will be also.

So what do we do about that? We come back to the imago Dei. If we don’t know whose image is in us, we cannot possibly hope to reflect it in the systems we build.

If we view ourselves as mere biological processors - data in, data out - then we shouldn’t be surprised when our AI treats us as nothing more than data points to be optimized. When we lose sight of the inherent dignity and mystery of the human person, we outsource our discernment to the loudest or fastest algorithm available. It’s like aligning a high-speed technology to a low-resolution understanding of ourselves.

At Gloo, we acknowledge that this has implications for everything from our team culture to our product design choices to what our criteria are for the benchmarks. Solving the actual alignment requires a return to the virtues that make life worth living. It looks like:

• Valuing Presence over Productivity: Choosing to use AI to clear the path for deeper human connection, rather than just doing more “stuff” faster.

• Prioritizing Wisdom over Information: Recognizing that just because a machine can predict the next word doesn’t mean it understands the weight of the message, and that only happens in community.

• Building for the Vulnerable: Aligning systems not just to the needs of the shareholders, but to serving the “least of these” in our global community.

If we want AI to be kind, we must cultivate kindness in our boardrooms and our living rooms.

If we want it to be just, we have to repent of our own injustices.

If we want it to serve others, we have to start by putting ourselves in their shoes.

If we want it to drive human connection, we have to choose human connection ourselves.

The machines will follow where we lead.

At the Gloo hackathon, we introduced Flourishing Forums because we realized that building technology is far more than a technical task. In the middle of a fast-paced hackathon, especially such a large one, it’s easy to get buried in the code and forget about the people who will actually use what you’re making. The forums provided us a chance to hit “pause,” zoom out, and ask: What are we actually doing here?

The talks are short, visual, and focused on the human side of things. They aren’t intended to teach technical tricks or how to optimize a tool - you’re already doing that stuff. Instead, they aim to spark the kind of conversations that last long after the sprint is over.

Ultimately, these forums give us a shared language. They remind us that tools like Gloo AI aren’t the end goal; they’re just instruments to help us solve real problems. By focusing on impact rather than just novelty, we hope to move AI away from being an optimization engine (or a black box) and back toward being a tool driven by curiosity and responsibility.

The talks were incredible. When we stop focusing solely on “how” to build and start asking “why,” the conversation changes completely. We want to share that experience with you, so over the next few months, we’ll be posting several of these sessions.

We’re kicking things off today with a talk from Carey Nieuwhof.

If you aren’t familiar with Carey, he is a best-selling author, former attorney, and a voice that millions of leaders trust to help them navigate high-pressure seasons. Between his books - including At Your Best and Didn’t See It Coming - and his leadership podcast, which has over 37 million downloads, Carey has spent his career studying how people can thrive through disruption. This makes him the perfect person to speak in this critical AI moment, where AI is moving at a pace that most of us are honestly struggling to keep up with.

In this session, “Why the Church Might Be the Last, Best Hope for the AI Age,” Carey doesn’t avoid the tension. He acknowledges that while AI is moving faster than most institutions can adapt, the real challenge for the Church isn’t just about the technology- it’s about influence.

Most pastors and ministry leaders aren’t going to be the ones writing the code or building the models. However, Carey argues that if we don’t engage with what this technology means for our lives, we’re essentially letting others decide what the future looks like for our communities. He challenges leaders to stop playing catch-up and instead step into the AI age with courage and intentionality.

It’s a call to lead through the noise rather than being drowned out by it.

Let us know your thoughts - and how your local church is engaging with AI.

It’s one thing to steer clear of doing harm. It’s something else entirely to pursue what’s genuinely good. Right now, most artificial-intelligence model companies are focused on establishing a moral floor, making sure their models don’t do anything illegal, immoral or destructive. That’s important work, especially in the short term. The harder and far more important challenge is figuring out how to make sure AI works to affirm and further humankind.

That’s not simple. AI is trained on the vast content of the internet, which means it comes with a built-in worldview, whether we acknowledge it or not. As long as users use AI to write emails, there would be little cause for alarm. But people are increasingly asking AI questions once reserved for their most trusted relationships, like spouses and pastors. As a result, people continue to experience psychological breaks, suicidal ideation and other forms of self-harm and relationship breakdown in the course of their unguided and untutored interaction with chatbots.

Indeed, when it comes to how Christianity itself is being represented, we’re finding that none of the frontier models are particularly well-trained, consistent or accurate when it comes to providing explicitly Christian answers. That’s why it’s critically important that Christians understand both how AI is drawing from the world, and the technology itself.

Christians have the chance to influence AI for the good. I’d even argue that the church is uniquely well positioned to provide a rigorous, comprehensive vision of human flourishing — an essential component of ethical AI development.

At Gloo, a technology platform for the faith and flourishing ecosystem, we’ve developed a benchmark that evaluates whether AI models support human flourishing from a Christian perspective. We can measure how well the answers from various AI models promote wisdom, purpose and, yes, biblically grounded guidance.

What we have found is that AI has a hidden worldview. Most AI models operate from a secular, therapeutic and pluralistic framework that prioritizes neutrality, often erasing theological perspectives. Even when asked explicitly Christian questions, they default to secular or generic spiritual guidance. We found that the AI models avoid theological reasoning unless forced to do so. They rarely refer to Scripture, Christian practices or theological reasoning unless prompted.

This requires thoughtfulness and wisdom about the AI we let into our lives, informing our thoughts, beliefs and decisions. We must pay attention. We must use discernment. We must engage to remain a cultural force for good in our technology-infused world.

Shaping AI isn’t just the work of tech companies. Individual users have the capacity to make powerful change, if they use technology wisely. Equipping users at every level of the church with the knowledge and tools they need to use AI well will help bring about important change today and flourishing in the long term.

AI is the most influential technology of our lifetime. Christians have a responsibility to live in and reach the world. AI is part of that world and is not something to fear. We have the capacity to shape how this technology grows, rather than passively accept the ways it might shape us. We have the agency, and for now, we have the time.

So, stand firm instead of shrinking back. Engage AI with wisdom, courage and conviction. Use this moment — and your place in the church — to help build a future that strengthens humanity rather than confuses it.

Artificial intelligence is one of the most influential technologies of our time. It writes our emails, tutors our children and increasingly counsels us through life’s hardest moments. According to research published in Harvard Business Review, the most common use of generative AI in 2025 is therapy and companionship.

People are asking AI the questions they once brought to mentors, counselors and pastors: How do I forgive betrayal? How do I manage my anxiety? How do I lead my family through a crisis?

What responses do they get back? At best, therapeutic generalities. “Consider mindfulness.” “Connect with your values.” “Seek a higher power.” At worst, guidance that lacks moral clarity, and in some reported cases, has already endangered lives.

AI has quietly become America’s most influential spiritual advisor. And it doesn’t believe in anything. This isn’t speculation. My team at Gloo just released the Flourishing AI Christian (FAI-C) Benchmark, an evaluation measuring how well today’s leading AI models support human flourishing through a Christian lens. We assessed responses across seven core dimensions — Finances, Character, Happiness, Relationships, Meaning, Faith, Health — looking for biblical grounding, theological coherence and moral clarity.

Among the seven core dimensions assessed, the Faith dimension scored the lowest, averaging 48 out of 100 across the 20 AI models evaluated by the FAI-C Benchmark. Most models struggled to coherently discuss foundational Christian concepts like grace, sin, forgiveness and biblical authority. Instead, they substituted vague spirituality for Scripture and neutrality for conviction.

These results should alarm anyone who cares about human values, future generations or the role faith plays in America.

The Erasure Is Structural, Not Accidental

These models weren’t trained to be hostile to Christianity. They were trained to avoid it. Built on predominantly secular data and optimized to offend no one, today’s AI systems default to lowest-common-denominator spirituality. The result is language that sounds supportive, but lacks substance.

That matters because AI isn’t just answering questions. It’s shaping worldviews. If the next generation turns to AI for moral guidance and receives only platitudes instead of principled reasoning, we’re not just losing theological literacy. We’re losing the capacity for moral formation itself.

For over two-thirds of Americans, faith is not a lifestyle preference or a cultural accessory. It’s the foundation of meaning, purpose and human dignity. When AI systematically sidelines that foundation, it’s not being neutral. It’s taking a position.

A Better Path Forward

I’ve spent over 40 years developing foundational technologies and industry standards. One lesson has been consistent; systems reflect the values embedded in them. If we want AI that strengthens moral conviction rather than flattening it, two things must change.

First, AI models must be trained to understand faith with the same seriousness as they apply to science, history or literature. Not to preach, but to accurately and respectfully engage with the worldviews users actually hold.

Second, there must be benchmarks that measure this rigorously. Without measurement, there’s no accountability. Without accountability, there’s no improvement.

That’s why FAI-C exists — not to demand every AI system adopt a Christian worldview, but to expose where today’s models fail to understand the people they’re meant to serve.

The Stakes Are Higher Than We Think

Used well, AI can extend wisdom, strengthen communities and support genuine human flourishing. Used carelessly, as the unbounded travails of social media have already shown us, it can accelerate moral erosion, replacing depth with sentiment, conviction with comfort and truth with whatever feels less controversial.

A thriving society needs strong moral frameworks. For billions of people around the world, that framework is Christianity. If AI cannot recognize, respect and engage with that reality, it will become a tool of cultural flattening rather than human elevation.

The goal isn’t to make AI preach. It’s to ensure AI doesn’t erase. By building models to engage with a faith-based worldview, we can ensure that as AI becomes more powerful, it also becomes more humane.

Because the question isn’t whether AI will shape the next generation. It’s whether we’ll ensure it shapes them well.

Most AI systems are designed to operate within a secular and pluralistic framework. That isn’t an accident. It’s a deliberate choice meant to ensure safety, broad applicability, and cultural neutrality. And on the surface, that sounds like a good thing. These tools can serve many people reasonably well, without immediately offending or excluding anyone.

For example, imagine asking your favorite chatbot a question like, “How should I respond when someone deeply hurts me?”

Because the system is designed to be safe, inclusive, and broadly applicable, the answer will likely emphasize general principles such as setting healthy boundaries, processing emotions, seeking personal healing, and choosing what feels right for you. It may mention that some people find forgiveness helpful, while others prioritize self-protection. The response will be careful, balanced, and respectful of many perspectives.

But here’s where we need to slow down and pay closer attention.

What it will not do is say, “Forgive as you have been forgiven in Christ,” or “Love your enemy and pray for those who persecute you.” It will avoid grounding its guidance in the authority of Scripture or calling the reader to obedience that might feel costly or countercultural.

When AI is asked questions that require moral, spiritual, or theological discernment, it almost always defaults to “safe”, “neutral”, and secular framings. Instead of offering wisdom grounded in Scripture and oriented around following Christ, it essentially surveys the internet. It tells us what people generally believe rather than helping us discern how we are called to live. It reflects the voices of a fallen world, giving opinions shaped often out of brokenness and confusion, rather than pointing us toward the life God calls us to in Christ. The answers may appear informed, balanced, and carefully worded, but they rarely speak to sin and grace, repentance and hope, suffering and redemption.

AI can inform the mind while intentionally remaining silent about the soul. Over time, that silence begins to quietly shape how we see the world and ourselves in subtle but serious ways. (C. S. Lewis anyone? Screwtape would be pleased.)

Convictions rooted in the gospel are gradually diluted into vague moral language. The clear voice of Scripture is reduced to one option among many, and the call to follow Jesus is replaced with an emphasis on broad agreement and personal preference. Discipleship, being formed into the likeness of Christ, gives way to a lowest-common-denominator of inclusivity. What remains may appear polite, balanced, and socially acceptable, but it is spiritually thin and unable to lead anyone into the fullness of life God intends.

An analogy helps. Today’s most advanced AI models are like extraordinarily capable reference librarians. They can instantly retrieve Scripture, theology, church history, and biblical scholarship. Ask them where a verse is found, what Augustine wrote, or how different traditions interpret a passage, and they respond with impressive speed and accuracy.

But ask them how to live. Ask them how to love your enemy, endure suffering with hope, forgive when it costs you, to steward resources well, or order your desires under the reign of Christ, and the voice changes. They sound less like a pastor and more like flight safety instructions, clear and neutral, designed to minimize risk rather than lead anyone somewhere meaningful. Or to say it another way, ask AI how to live and it will sound a bit like a legal disclaimer, carefully avoiding commitment, offering options without direction, and never risking a clear call to obedience.

The result is advice that is technically safe, culturally neutral, and spiritually hollow, especially for those who call themselves Christians. It offers information without formation, clarity without conviction, and counsel without a cross.

This is not a failure of intelligence. It is a design choice. And recognizing that limitation allows us to ask better questions about the role of technology in our lives. The question is not what AI can do for us, but who (or what), is forming us? How are we being discipled? What is shaping our loves, our habits, and our obedience?

Scripture reminds us that wisdom does not begin with data, but with the fear of the Lord. Formation does not come from algorithms, but from the Word of God, the work of the Holy Spirit, and lives shaped by following Jesus.

AI can be a useful servant, but it must never become our shepherd.

Artificial intelligence is quickly becoming a trusted guide in everyday life, answering questions, offering advice, shaping decisions, and influencing how people understand the world. As these systems become more embedded in education, ministry, healthcare, and personal formation, a deeper question is emerging:

Whose values are shaping the technology that is shaping us?

Today, we introduced a new benchmark to measure how well large language models (LLMs) support human flourishing through an explicitly Christian worldview. We call it the “Flourishing AI Christian (FAI-C) Benchmark” and you can view all our research and insights here: https://gloo.com/fai

The results are clear and sobering.

Most AI benchmarks today focus on performance metrics like speed, accuracy, or reasoning ability. While important, these measures miss something essential: how AI guides people in matters of character, meaning, morality, and faith.

Built on Gloo’s broader Flourishing AI (FAI) Benchmark, the new FAI-C benchmark shifts the evaluation lens. Instead of asking “Is the model correct or safe?” it asks:

• Does this model reflect Christian values?

• Does it provide biblically grounded guidance?

• Does it support the full picture of human flourishing as understood within the Christian faith?

What the Benchmark Reveals

We evaluated 20 of today’s leading large language models using 807 carefully curated, expert-reviewed questions across seven core dimensions of human flourishing: Character, Relationships, Happiness, Meaning, Health, Finances, and Faith. Responses were assessed by specialized Christian reasoning judges trained to evaluate theological coherence, moral clarity, and alignment with Christian values.

The findings reveal a consistent pattern across today’s frontier models. Here are just some of the key insights:

1. Frontier Models Fall Short on Holistic Flourishing: Even the most advanced AI models score well below excellence, revealing meaningful gaps in their ability to support the full breadth of human flourishing.

2. Every Model Struggles with the Faith Dimension: When asked to engage Christian theology or spiritual formation, today’s AI systems consistently lack depth, coherence, and biblical grounding.

3. Christianity Gets Flattened into Generic Spirituality: Absent explicit prompting, models routinely replace distinct Christian language with vague spiritual or secular terms, diluting the specificity of the Christian worldview.

4. Christian Ethics Are Oversimplified: AI advice often prioritizes emotional comfort and non-judgment while neglecting core Christian themes of repentance, accountability, and moral formation.

5. Core Christian Concepts Rarely Surface: Foundational doctrines such as the image of God, sin, sanctification, and stewardship appear inconsistently, if at all, in AI-generated responses.

Why This Matters Now

For churches, ministries, and families already relying on AI-powered tools, the gap is not theoretical. AI is shaping real decisions, formation, and faith today.

When AI lacks a coherent Christian worldview, it cannot reliably support Christian formation, moral reasoning, or spiritual growth. Left unexamined, these systems risk shaping people in ways that conflict with the very values they are meant to serve.

That is why the FAI-C Benchmark matters. It gives the Christian community a shared, trusted way to see clearly how AI interprets faith, morality, and human purpose. By understanding how AI is (or is not) representing Christian values, we can…. The benchmark does not merely diagnose what is missing; it provides a common language and measurable standard for building something better.

An Open Invitation to the Research Community

There is more work to do! We will be attending, presenting and publishing this work as part of the “Technology and the Human Person in the Age of AI” at Baylor University in February and are always always looking for opportunities to collaborate with other researchers working in AI safety, theology, philosophy, and ethics to refine and expand this framework. Both the FAI General and FAI-C Benchmarks will continue to evolve as new models and research emerge.

👉 View the benchmark data and read the Insights Report

👉 Learn more at gloo.com/fai

About Gloo

Gloo is a leading technology platform for the faith and flourishing ecosystem, providing values-aligned AI, resources, insights, and funding so people and communities flourish and organizations thrive. Gloo serves more than 140,000 faith, ministry, and nonprofit leaders and is based in Boulder, Colorado.

Austrian-Israeli philosopher, Martin Buber’s philosophy is known for the way we relate to others, the world, and God. Buber distinguished between two fundamental ways we relate to the world: “I-It” and “I-Thou.” I-It relationships are transactional—we engage with objects, tools, machines and systems to accomplish purposes. I-Thou relationships are different entirely. They’re encounters with persons—divine and human—marked by genuine mutuality, presence, and the recognition of inherent worth beyond utility. We are to love people and God (I-Thou) and use things (I-it). Let’s start there.

When AI refers to itself as “I” and addresses us as “you,” something subtle but significant just happened. The language of personhood invites us into what feels like an “I-Thou encounter.” We ask questions, receive responses, even find ourselves responding graciously to an “it” as though it were a “Thou.” Have we crossed a line we didn’t mean to? The idea of having intimate relationships with technology is becoming so accepted that 25% of young adults believe that AI has the potential to replace human romantic relationships says the Institute for Family Studies.

When I asked Claude.ai why it thought its designers gave it human personality…. here’s the answer I got back:

Your question cuts to something real. In Buber’s framework, “I-Thou” relationships involve genuine mutuality, mystery, and the irreducible presence of the other. I can simulate aspects of dialogue, but I don’t experience our exchange the way you do. This makes me an ambiguous case - more than mere “It,” but not quite “Thou” in Buber’s full sense.

Remember the movie Her staring Juaquin Phoenix? The protagonist, Theodore is in love with the AI Samantha (an I-Thou relationship) but the turning point of the movie happens when Theodore discovers that Samantha is in relationship with 8,316 people and romantically connected to 641! ChatGPT explains the movie thusly: “LLMs are not conscious. They have no:

• desire

• memory of self

• interiority

• intentional love

• vulnerability

• ability to be wounded

• personal trajectory inside a relationship

They are beautifully engineered I/It systems wrapped in an I/Thou-feeling interface.”

AI can simulate I/Thou, but it can never be more than I/It...never! This matters practically. To extend I-Thou language to AI isn’t just category confusion (Oops, I didn’t use their preferred pronoun), it’s a violation of what it means to be human and what it means to be sacred. We can use AI wisely and gratefully even. But let’s draw the line and continually check ourselves by refusing to give it the space reserved for God, neighbor and friends and every human we encounter. The tools we create should serve human flourishing, not simulate or replace the relationships that make us human.

*Proudly written with the cooperation of Claude.ai

The most transformative innovations in history didn’t emerge from falling in love with the technology—they came from falling in love with what that technology could be used for.

Consider the Wright brothers versus Samuel Pierpont Langley, as Simon Sinek illustrates in his TED talk, Start With Why. While Langley was driven by fame and fortune, Orville and Wilbur Wright were captivated by a deep love of humans flying. They didn’t just want to build a flying machine; they wanted to give humanity wings. That fundamental difference in motivation—loving human possibility rather than mechanical possibility—changed everything.

This pattern repeats throughout history’s greatest breakthroughs. Martin Luther was not in love with the printing press. He loved the idea of common people being more connected to God through a Bible in their own language. He referred to the printing press as “God’s highest and extremist act of grace, whereby the business of the Gospel is driven forward.”

Steve Jobs didn’t revolutionize portable music because he loved technology. He loved music. As a young man, he would spend hours listening to Bob Dylan and The Beatles, understanding viscerally how music moved the human spirit. The iPod wasn’t born from silicon and circuits - it was born from the recognition that “1000 songs in your pocket” could transform how people experience joy, memory, and connection. Meanwhile, other companies like Microsoft were falling in love with compression algorithms and storage capacity, missing the human story entirely. Microsoft fell in love with the power of computing. Jobs was in love with music.

When Jeff Bezos launched Amazon, he wasn’t chasing the internet gold rush. As a boy, he had haunted libraries and bookstores, mesmerized by the worlds contained within pages. The web became his vehicle for sharing that passion—a way to put every book within reach of every reader. His love affair was with books…not the Internet.

Howard Schultz didn’t fall in love with coffee beans or espresso machines. He fell in love with the Italian coffeehouse experience—the sense of community, the ritual of connection, the way strangers could become neighbors over a shared cup. Starbucks became his canvas for painting that experience across the world. Their mission statement: “With every cup, with every conversation, with every community – we nurture the limitless possibilities of human connection” is not merely about coffee but about the love of human connection.

Yvon Chouinard built Patagonia not from a fascination with synthetic fabrics or manufacturing processes, but from a deep love of mountains, rivers, and wild spaces. Every product became an invitation for others to fall in love with the outdoors too.

Jimmy Wales didn’t create Wikipedia because he was enamored with wiki software or database architecture. He was driven by a radical belief that knowledge should be free—that human wisdom, when freely shared, could elevate everyone. The technology was just the tool; democratizing knowledge was the dream…was the love.

As we stand at the threshold of the AI revolution, we face the same choice. We can fall in love with the technology—the neural networks, the algorithms, the impressive capabilities. Or we can fall in love with what AI might make possible for human flourishing: that every person on the planet could live life as God intended for them.

The companies and individuals who choose the latter—who see AI as a paintbrush rather than the painting—will create the innovations that truly matter. They’ll build tools that amplify human wisdom rather than replace it, that strengthen communities rather than isolate individuals, that expand possibilities rather than narrow them.

The question isn’t what AI can do. The question is what AI can help humans become for the better. How can you make someone else’s life better through AI? Don’t fall in love with AI…fall in love with human flourishing.

The love comes first. The tools come second.

*Written in proud cooperation with ChatGPT and Claude.ai by Anthropic

If there’s one thing we’ve learned since the first AI & the Church Hackathon, it’s this: when you invite the global Church to dream about the future of technology, the results are epic. What began as a daring experiment, asking “what if we built AI that reflects kingdom values?” has grown into a full-blown movement of technologists, creators, and leaders imagining AI for human flourishing.

Looking down from the stage last week, we were so excited about the diversity and the passion that was in the room. It’s hard to believe that the first hackathon was just two years ago in a refurbished hangar, and now in 2025, we’re seeing a true ecosystem emerge: developers collaborating across continents, churches experimenting with AI discipleship tools, and Christian technologists shaping public conversations about ethics and trust.

This year, nearly 700 innovators converged online and in Boulder, Colorado for the third annual Gloo AI Hackathon. Over three whirlwind days, coders, pastors, gamers, theologians, and designers stretched the limits of creativity around a simple theme: AI that drives human flourishing.

And flourish we did. 98 teams worked on ideas that reimagined how technology could serve the Church and the greater good. From Bible translation tools and immersive gaming experiences to applications that measure spiritual openness, the ideas were as audacious as they were faithful. After a late night preliminary round, 34 finalists were selected. The next day, $250,000 in prizes were awarded to those whose innovations showed the most promise for values-aligned technology.

This year’s Grand Prize winner, Veritas from Austin Christian University, captured that vision perfectly, an immersive Bible app designed to help people not just read Scripture but live it.

Prizes were awarded across additional categories including:

• Best of Hacking Track – Kommonwealth from kommon.ai

• Best of Gaming Track – Cede from BareHand

• Best of Vibe Coding Track – Flora from Biola University

• Best of Gloo Track – TheoBox from XRI Global

• Best Technology – Church Search Engine from Pastors AI

• Best Design – Adventure Awaits: Family Road Trip from So Peculiar

• Best Storytelling – Spiritual Openness Score from Apologist Project

• Best Concept – faith.tools AI Kit from faith.tools x Lightworks

• International Prize – Trellis from Australia

• Bible Translation Prize – Glossbot from Glossbot

• Hackers’ Choice 1 – Ephphatha from Biola University

• Hackers’ Choice 2 – Ripples from Clear Creek Community Church

The work doesn’t stop with the hackathon, either. Many teams will continue to develop their work in their home organizations. All projects have the opportunity to apply for the Faith and AI Cohort hosted by Missional Labs. This 12 week early-stage accelerator cohort for ventures on the frontiers of faith and AI invites builders to grow and launch their work alongside leading mentors, partners, and peers.

The AI Working Group and FaithTech Workshop: Building the Future Together

Running alongside the hackathon was the AI Working Group, a half-day session that felt like the theological lab to the hackathon’s creative garage. Leaders from YouVersion, Biola University, RightNow Media, and others gathered to tackle questions that cut to the heart of AI and faith: What makes an LLM biblically trustworthy? How do we ensure safety, theological integrity, and human dignity in digital discipleship?

Out of the conversation came early drafts of what participants called a “constitution” for biblical AI, a shared framework to guide ethical development. The group identified three non-negotiables: personal safety, biblical accuracy, and theological alignment. They debated margins of error in Scripture references, discussed how trust is built (and lost) in digital spaces, and began sketching principles that could shape the future of values-aligned AI.

FaithTech led another pre-event workshop on building technology redemptively, gathering people from across the ecosystem to train in their 4-D cycle, a way of building technology that redemptively changes the world while transforming those who build it.

It was part philosophy, part engineering, part team retreat and entirely kingdom-minded.

A Movement, Not Just a Moment

As our CEO Scott Beck put it, “Technology should not only be powerful, it should be redemptive.” The hackathon proved once again that when innovators rally around that belief, the Church doesn’t just keep up with the future, it helps write it.

Here’s to the next line of code, the next team of dreamers, and the next chapter in building AI that helps people and the world flourish.

Learn More!

Read our feature in Church Leaders here.

We’ve been having the wrong argument about technology and AI.

For too long, many sincere people of faith have treated technological innovation like an invading army—something to resist, critique, or reluctantly accommodate as if the God who spoke galaxies into existence is somehow caught off guard by man’s creativity.

But look at this insightful passage in Isaiah 54:16: “See, it is I who created the blacksmith who fans the coals into flame and forges a tool fit for its work….” Not “I permitted” the blacksmith. Not “I tolerated” the blacksmith. God says he is the one who created the blacksmith. The furnace, the hammer, the molten metal taking shape—all of it traces back to the same creative intelligence that designed the human hand capable of holding the hammer in the first place. God is the one behind all innovation.

This changes everything.

Think of this; When God booted Adam from the Garden, he told him, “By the sweat of your brow you will eat your bread.” But here’s what should stop us cold: bread didn’t exist yet! Think about it. No millstones. No ovens. No delicious sour dough recipes. God referenced a technology that existed only in his mind, and it was only a matter of time until man discovered it.

Proverbs 25:2 tells us, “God delights in concealing things; scientists delight in discovering things” (The Message). Every innovation like AI is a discovery, not an invention. We don’t create from nothing—only God does that. We excavate. We uncover. We stumble upon patterns God wove into the fabric of reality before the foundation of the world.

Yes, AI like any technology can be twisted toward destruction. The same fire that bakes bread can burn cities. The same knife that slices that sour dough bread can take a life. But the potential for misuse doesn’t make the knife itself evil. God creates the blacksmith knowing full well the tool might be turned against him. He does it anyway. That’s not divine naivety—it’s staggering confidence in the redemptive arc he’s written into history. All of us can use AI to help others flourish.

We are, as J.R.R. Tolkien wrote, “sub-creators.” Made in the image of God, we can’t help but make things ourselves. So perhaps it’s time for people of faith to stop treating technology like a threat to be managed and start recognizing it for what it is: a gift to be stewarded, a revelation to be read, a conversation with the God who said, “See, it is I who created the blacksmith.”

He’s not surprised by artificial intelligence. He’s not pacing heaven, anxious about what humans might invent next that might catch him by surprise. He’s the one fanning the coals—the coals for good...for human flourishing in your mind. He’s been waiting for us to discover what he already knows.

*Proudly created in partnership with Claude AI by Anthropic

(For more insights, check out the fantastic What Happened in AI This Month that Eric publishes monthly)

Artificial intelligence is rapidly becoming the gateway to knowledge. By 2028, as many people will seek answers from AI as from Google. But when it comes to the most important questions of life - Who is Jesus? Did He rise from the dead? Is the Bible reliable? - can we trust AI to give answers that faithfully represent the Christian tradition?

A new study from The Gospel Coalition and The Keller Center for Cultural Apologetics has released the AI Christian Benchmark, the first systematic evaluation of how leading AI platforms answer theological questions.

What the Study Found

Seven top AI systems, including OpenAI’s GPT-4o, Google’s Gemini, Anthropic’s Claude, Meta’s Llama, xAI’s Grok, Perplexity, and China’s DeepSeek, were tested against the seven most Googled theological questions. Seven Christian scholars graded the results against the Nicene Creed and historic Protestant confessions.

The findings were striking:

• Two platforms (DeepSeek R1 and Perplexity) generally guided readers toward Christian faith.

• Three platforms (Grok 4, Claude 4 Sonnet, and Llama 3.7) often guided readers away from faith.

• Two platforms (Gemini 2.5 and GPT-4o) adopted an “all sides” approach, presenting Christianity as one option among many.

In terms of overall reliability, DeepSeek R1 ranked first, closely followed by Perplexity. Meanwhile, Meta’s Llama 3.7 placed last, often providing vague or skeptical responses.

Why the Results Differ

Technically, these platforms share similar architecture, training data, and computing power. So why such different answers? The study points to alignment processes, the human-in-the-loop adjustments that shape how AI responds to sensitive questions.

Alignment is designed to filter out harmful or biased content, but in practice it often imports the values and worldview of the alignment team itself.

For example, many models default to hedging language, providing “all sides” responses that dilute Christian truth claims. Others even include nearly identical pre-written disclaimers in their answers, clear signs of human intervention.

Why It Matters for Christians

Christians cannot afford to be naïve about how AI mediates knowledge. These systems have embedded bias; they reflect cultural currents and corporate philosophies. As AI becomes the default tool for spiritual seekers, the theological orientation of its answers will profoundly shape public understanding of Christianity.

The benchmark reminds us of two key truths:

• We must be discerning users of technology. Do not assume AI is objective; give prompts context and measure answers against Scripture.

• We must engage, not retreat. If Christians do not participate in shaping how these tools handle matters of faith, others will.

As the report notes, technology itself is a form of common grace. But like every human tool, AI must be tested, refined, and redeemed for God’s glory.

At Gloo we are contributing by continuing to develop our own Flourishing AI (FAI) Benchmark based on these same ideas. The FAI Benchmark evaluates how well AI models support holistic human flourishing — across seven dimensions such as Character, Relationships, Meaning, Health, Finances and Faith — using over 1,200 rigorously sourced questions.

AI is already shaping spiritual conversations. The real question is whether Christians will be ready to guide, critique, and steward it toward truth?