Blog

Posted on

Jensen Huang’s Sovereign AI and the Call to Digital Discipleship: Why Nations Need More Than Computing Power

Jensen Huang’s latest push centers on “sovereign AI” — the idea that nations need their own AI infrastructure, data, and models to maintain digital independence. Speaking at recent conferences, Huang argues that countries must build local AI capabilities rather than depend entirely on foreign systems, combining their unique cultural knowledge with computing power to create AI that serves their specific populations.¹

The concept resonates beyond geopolitics. It’s fundamentally about stewardship — who controls the tools that shape how people access information, make decisions, and understand their world.

The Great Commission Requires Local Infrastructure

When Jesus commissioned his followers to “go and make disciples of all nations” (Matthew 28:19, ESV), he wasn’t envisioning a centralized Jerusalem-based operation. The early church spread through local communities, adapting the gospel message to different cultures while maintaining core theological truth.

Huang’s sovereign AI framework mirrors this pattern. Just as the gospel needed local expression, Paul writing differently to Romans than to Corinthians, digital discipleship requires infrastructure that understands local context.

At Bible Gateway, we see this daily. Our 200+ translations serve 70+ languages precisely because discipleship isn’t one-size-fits-all. A believer in Chennai needs Tamil commentary. A pastor in São Paulo needs Portuguese study tools. A seminary student in Seoul needs Korean cross-references.

But here’s where Huang’s vision gets complicated for faith communities: sovereignty over AI systems means sovereignty over interpretation. When algorithms shape how Scripture gets searched, studied, and understood, the question becomes: who is training those models?

The Stewardship Problem Hidden in Infrastructure

“Whoever is faithful in very little is also faithful in much” (Luke 16:10, ESV). This verse cuts to the heart of why sovereign AI matters for Christian organizations.

Every search ranking, every recommendation algorithm, every content filter represents a micro-decision about what matters most. At Bible Gateway, when someone searches “love,” do we surface 1 Corinthians 13, John 3:16, or Romans 8:38-39 first? The algorithm makes that choice thousands of times daily across millions of users.

Right now, most faith-based platforms depend on external AI systems like Google’s search algorithms, Amazon’s cloud infrastructure, OpenAI’s language models. We’re essentially outsourcing discipleship decisions to secular systems trained on secular priorities.

That’s not necessarily wrong. But it’s worth examining.

What Sovereign AI Looks Like in Practice

Here’s where Huang’s framework gets practical for faith tech builders. Sovereign AI doesn’t require building everything from scratch, it requires intentional control over the pieces that matter most.

For Bible Gateway, that might mean:

  • Training language models on theological texts, not just Wikipedia
  • Building search algorithms that understand scriptural context, not just keyword matching
  • Creating recommendation engines that prioritize spiritual growth over engagement metrics

I’m not advocating for Christian-only AI systems. The gospel spreads through engagement with the broader world. But I am suggesting we need infrastructure designed with discipleship as a first-class concern.

Consider our reading plan completion rates. When we launched plans optimized by secular engagement algorithms, completion dropped after Day 7 then users got recommendations that prioritized “interesting” content over spiritual discipline. When we rebuilt the system around formation rather than retention, completion improved 23% over six months.

The difference wasn’t the technology. It was the training data and optimization targets.

The Wisdom of Solomon Applied to AI Infrastructure

“The simple believe everything, but the prudent give thought to their steps” (Proverbs 14:15, ESV). Solomon’s wisdom about discernment applies directly to how we build AI systems.

Huang’s sovereign AI concept recognizes that different communities need different approaches to intelligence. A financial AI system optimized for Wall Street trading won’t serve a rural credit union. A healthcare AI trained on urban hospital data won’t understand rural clinic challenges.

Similarly, AI systems trained on secular content patterns won’t naturally understand spiritual formation needs. When Ethan Mollick talks about co-intelligence, he’s describing partnership between humans and AI. But what kind of partnership do we want for discipleship?

At HarperCollins Christian Publishing, we’re starting to answer that question. Not by building competing AI infrastructure, we don’t have Google’s resources, but by curating the training data and fine-tuning the outputs for spiritual formation. This is what GLOO was doing when I worked there.

Building Digital Discipleship Infrastructure

The practical implementation isn’t about creating Christian ChatGPT. It’s about ensuring the tools that shape faith formation are built with discipleship in mind.

Three areas where this matters most:

Search and Discovery: When someone searches “suffering” in our Bible study tools, do they get academic theology or pastoral care? Both have value, but the algorithm’s choice shapes the user’s spiritual journey.

Content Recommendations: Our reading plans serve 23 million annual users.² Every “what to read next” suggestion influences someone’s Bible engagement. Training those systems on spiritual formation research rather than generic engagement metrics changes everything.

Translation and Commentary: As AI-assisted translation tools proliferate, who’s ensuring theological accuracy? When AI comes for sermon prep, pastors need tools trained on sound doctrine, not just persuasive rhetoric.

The Cost of Digital Dependence

“Do not put your trust in princes, in human beings, who cannot save” (Psalm 146:3, ESV). This psalm warns against depending entirely on external powers whether political or technological.

Huang’s sovereign AI argument recognizes that complete dependence on foreign AI systems creates vulnerability. For nations, that might mean security risks. For faith communities, it might mean theological drift.

I’m not advocating for technological isolationism. The global church benefits from shared tools and resources. But I am suggesting we need more intentionality about where our digital discipleship infrastructure comes from and how it gets trained.

Search algorithms, content curation, and user experience design are the pieces that directly shape spiritual formation and need to be understood and protected. Not because we’re better engineers, but because discipleship is our primary mission.

The Path Forward

Huang’s sovereign AI vision offers a framework, not a blueprint. For Christian product builders, the question isn’t whether to build competing infrastructure, it’s how to maintain faithful stewardship over the tools that shape discipleship.

That might mean:

  • Partnering with AI providers who understand faith-based applications
  • Investing in fine-tuning and training data that reflects theological priorities
  • Building internal capabilities for the functions that most directly impact spiritual formation
  • Creating open-source tools that serve the broader faith community

The Tower of Babel reminds us that technology without wisdom leads to confusion. Huang’s sovereign AI concept,  adapted for faith communities, offers a path toward digital discipleship that serves spiritual formation rather than just technological efficiency.

The question for Christian product leaders: Are we building tools that make disciples, or are we just building tools?

¹ Jensen Huang, keynote address at COMPUTEX 2024: “Sovereign AI and the Future of Computing”
² Bible Gateway internal analytics, 2024 annual reading plan enrollment data

Photo by Avesta on Unsplash

Posted on

John 21:5-6 and the Art of Asking Better Questions: Why AI Prompting Is Like Jesus Teaching His Disciples to Fish

“Then Jesus called out to them, ‘Friends, haven’t you any fish?’ ‘No,’ they answered. He said, ‘Throw your net on the right side of the boat and you will find some.’ When they did, they were unable to haul the net in because of the large number of fish.” (John 21:5-6, NIV)

The disciples had been fishing all night with nothing to show for it. Then Jesus, who they didn’t immediately recognize, asked one simple question that changed everything. Not “Why aren’t you catching fish?” or “Have you tried different bait?” Just: “Haven’t you any fish?”

That question led to instruction. The instruction led to abundance.

When someone struggles with AI prompting, they’re casting their nets over and over, getting frustrated with empty results, convinced the tool is broken. But like the disciples, they’re often fishing in the wrong spot with the wrong approach.

The art isn’t in the casting, it’s in learning to ask better questions and knowing where to throw the net. Obviously, the disciples knew how to fish and this story isn’t really about fishing, it’s about obedience and trust, but I’m trying to use a metaphor and I’m not really that good at them.

The Problem With Most AI Interactions

I see this pattern constantly. Users approach AI tools the same way they approach search engines: throw in some keywords and hope for the best. But AI isn’t Google. It’s more like a really smart intern who needs context, direction, and clear expectations.

The disciples were experienced fishermen. They knew how to cast nets, repair equipment, read weather patterns. Most people struggling with AI aren’t lacking technical skills, they’re lacking the right framing.

Jesus didn’t give them a fishing tutorial. He asked a diagnostic question, then provided specific direction: “Throw your net on the right side of the boat.”

That specificity matters. “Right side” isn’t arbitrary, it’s based on understanding conditions they couldn’t see from their position in the boat. Jesus had a vantage point they didn’t.

The Anatomy of Better Questions

When I work with teams on AI integration for sermon prep, the breakthrough moment isn’t technical. It’s when they stop asking “How do I make AI write my sermon?” and start asking “How do I help AI understand my congregation’s needs?”

The difference:

Fishing in the wrong spot: “Write me a sermon on forgiveness.”

Throwing the net on the right side: “I’m preaching to a congregation that’s 60% over 50, many dealing with family estrangement after the 2020 election divisions. They’re tired of political sermons but need biblical hope for restoration. Help me write a 20-minute sermon on forgiveness that acknowledges real hurt without being preachy, using Matthew 18:21-22 as the primary text, with two personal application points they can act on this week.”

The second prompt gives AI the context it needs to be helpful. Like Jesus with the disciples, it provides specific direction based on understanding the full situation.

Why This Matters for Digital Discipleship

The disciples’ empty nets weren’t just about breakfast. John tells us this story in the context of restoration, Peter’s reinstatement, the commissioning to “feed my sheep,” the establishment of early church leadership. The fishing miracle was functional, but it served a larger discipleship purpose.

AI in ministry works the same way. The technical capability (generating text, analyzing data, creating content) serves the larger mission of discipleship. But like the disciples, we need to learn where to cast the net.

At Bible Gateway, we’re seeing this play out with 23 million monthly users across 200+ translations. The users who get the most value aren’t necessarily the most technically sophisticated — they’re the ones who understand how to frame their spiritual questions in ways that digital tools can support.

A user searching “hope Bible verses” gets generic results. A user searching “Bible verses about hope after job loss, specifically for someone who feels God has abandoned them” gets targeted, actionable content that can actually help with discipleship.

The difference isn’t in the search technology, it’s in learning to ask better questions.

The Jesus Method of AI Prompting

Jesus’s interaction with the disciples gives us a framework for effective AI engagement:

Start with diagnosis. “Haven’t you any fish?” establishes the current state. Before jumping into solutions, AI needs to understand what you’re actually trying to accomplish. Not just the task, but the context around it.

Provide specific direction. “Throw your net on the right side” isn’t vague inspiration. It’s actionable guidance based on understanding the full situation. Good AI prompts are similarly specific about desired output, tone, length, audience, and constraints.

Trust the process. The disciples could have argued about which side of the boat was better. Instead, they followed the instruction. AI works best when you iterate based on results, not when you debate the approach.

Recognize the bigger picture. This wasn’t really about fishing, it was about discipleship. Using AI like this isn’t really about efficiency, it’s about enabling better ministry, better products, better service to people who need what you’re building.

Practical Applications for Ministry and Product

This principle scales across everything I work on. Whether it’s helping pastors with AI sermon preparation or building features for Bible Gateway’s global user base, the pattern holds: better questions lead to better outcomes.

For pastors: Instead of asking AI to “help with Bible study preparation,” try: “I’m teaching a small group of new believers, mostly in their 20s and 30s, about spiritual disciplines. They’re interested but overwhelmed by traditional approaches. Help me design a 4-week study on prayer that feels accessible and practical, with weekly exercises they can actually complete.”

For product teams: Instead of asking AI to “analyze user feedback,” try: “Review these 200 support tickets from the past month. Our mobile app’s Bible reading plans have a 40% completion rate, but we don’t know why people drop off. Identify patterns in user complaints that might indicate specific friction points in the first two weeks of plan usage.”

The difference is specificity informed by context, which is exactly what Jesus provided the disciples.

Why the Right Side of the Boat Matters

The disciples caught so many fish they couldn’t haul the net in. Not because the fish suddenly appeared, but because they were fishing where the fish actually were.

In the wisdom tradition, this is about alignment and understanding how things actually work rather than how we think they should work. AI isn’t magic, but it is powerful when applied with wisdom and clear direction.

The abundance wasn’t in the tool (the net) or even the technique (the casting). It was in the guidance that led them to the right place at the right time with the right approach.

For those of us building digital discipleship tools, this matters enormously. We’re not just solving technical problems, we’re helping people encounter God through technology. The quality of that encounter often depends on learning to ask better questions.

Sermon Illustration

The disciples had been fishing all night with empty nets. They knew how to fish — they were professionals. But when Jesus asked if they had caught anything and told them to throw their net on the right side of the boat, everything changed. Suddenly they caught so many fish they couldn’t pull the net in.

Sometimes our prayers feel like those empty nets. We’re asking God for help, but we’re not seeing results. Maybe the issue isn’t God’s willingness to provide, maybe it’s learning to ask better questions. Instead of “God, help me,” try “God, help me understand what You want me to learn through this situation.” Instead of “God, fix this,” try “God, show me how to respond faithfully right here.” The abundance might not be in getting what we think we want, but in learning to ask for what we actually need. And like the disciples, we might discover that the breakthrough was there all along, we just needed better direction about where to cast our nets.

Photo by Ankit Manoharan on Unsplash

Posted on

Ethan Mollick’s Co-Intelligence and the Biblical Call to Wisdom: Why AI Partnership Requires More Than Technical Skill

Ethan Mollick, co-director of Wharton’s Mack Institute for Innovation Management, has spent the last two years making a compelling case that we’re entering an era of “co-intelligence” — where humans and AI work together as cognitive partners rather than in a traditional tool-user relationship.¹ His core thesis: the most productive future isn’t human replacement by AI, but human augmentation through AI, where both parties contribute complementary strengths to problems neither could solve alone. This partnership model, Mollick argues, requires us to develop entirely new skills around delegation, collaboration, and what he calls “cyborg” thinking.

As someone building products for millions of users, I keep coming back to a question Mollick doesn’t directly address: if AI is becoming our cognitive partner, what does wisdom look like in that partnership?

The answer, I think, starts in Proverbs.

The Wisdom Literature Has Something to Say About AI Partners

“Plans fail for lack of counsel, but with many advisers they succeed.” (Proverbs 15:22, NIV)

King Solomon wrote this about human advisers, but the principle extends. The Hebrew word for “counsel” here is sod — it means not just advice, but the kind of intimate consultation that comes from deep understanding of both the problem and the person facing it. It’s the difference between getting information and getting wisdom.

Mollick’s co-intelligence framework captures something biblical that most AI discussions miss: partnership requires discernment about what each party brings. In my daily work, I’ve watched this play out in real time.

When my team started experimenting with AI-assisted content curation, the first instinct was pure efficiency — let the AI scan, categorize, and recommend. Classic tool thinking. The results were technically accurate but spiritually hollow. AI could identify themes in Scripture but couldn’t discern why Romans 8:28 resonates differently for someone walking through grief versus someone making a career change.

The breakthrough came when we shifted to what Mollick would recognize as co-intelligence: AI handling pattern recognition across millions of reading behaviors while humans provided the pastoral wisdom about what those patterns actually meant for individual spiritual formation.

What Co-Intelligence Looks Like in Faith Tech

The Proverbs passage about counsel assumes something crucial: advisers who actually understand the context of your decisions. This is where most AI implementations in faith contexts fall short — not because the AI lacks capability, but because we haven’t thought carefully about what wisdom requires.

“The simple believe anything, but the prudent give thought to their steps.” (Proverbs 14:15, NIV)

Applied to AI partnership, this verse cuts both ways. We can’t be “simple” about what AI tells us, but we also can’t be prudent if we’re trying to solve everything ourselves.

This looks like AI identifying reading patterns — which passages get highlighted most, where people stop in reading plans, which search terms spike during cultural events. But the decision about what those patterns mean for product design? That requires human discernment informed by pastoral experience, theological training, and understanding of how spiritual formation actually works.

Mollick talks about this as “keeping humans in the loop,” but I’d frame it differently: keeping wisdom in the loop. The goal isn’t human involvement for its own sake — it’s ensuring that the partnership produces something that serves human flourishing, not just human efficiency.

The Delegation Problem: More Than Task Management

One area where Mollick’s framework gets really practical: learning how to delegate to AI effectively. This isn’t just about prompt engineering, it’s about understanding what kinds of problems benefit from AI’s strengths (pattern recognition, rapid iteration, handling scale) versus what needs human judgment (context interpretation, ethical reasoning, spiritual discernment).

“Commit to the Lord whatever you do, and he will establish your plans.” (Proverbs 16:3, NIV)

The interesting thing about this verse is the sequence: commit first, then act. In AI delegation, we often reverse this, we act first (deploy the AI solution) and try to align it with our values later.

I’ve been thinking about this in the context of sermon preparation tools. AI is definitely coming for sermon prep, and the early products are impressive from a technical standpoint. But most of them are solving the wrong problem by optimizing for content generation rather than spiritual formation.

A co-intelligence approach would start with the theological question: what’s the actual purpose of sermon preparation? Is it to produce content, or is it to help pastors engage deeply with Scripture so they can shepherd their congregations more effectively?

If it’s the latter (and I think it is), then AI partnership looks different. AI handles the research heavy lifting of cross-referencing commentaries, identifying thematic connections, surfacing relevant cultural context. The pastor handles the spiritual discernment of understanding their congregation’s specific needs, wrestling with how the text speaks to current circumstances, crafting application that connects eternal truth to daily life.

The Stewardship Question

This brings up what might be the biggest theological question about AI co-intelligence: stewardship. If we’re called to be faithful stewards of the gifts and resources God gives us, what does faithfulness look like when one of those resources is artificial intelligence?

“From everyone who has been given much, much will be demanded; and from the one who has been entrusted with much, much more will be asked.” (Luke 12:48, NIV)

AI capability definitely falls into the “much has been given” category. The question is what “much will be demanded” looks like in practice.

Mollick’s work suggests we’re still in the early stages of figuring this out. His research at Wharton shows that even sophisticated knowledge workers are using AI at maybe 20% of its potential, mostly because we’re still thinking about it as an advanced search engine rather than a cognitive partner.

But I wonder if that’s actually wise restraint rather than missed opportunity. The Tower of Babel was fundamentally about the misuse of technological capability, not technology itself, but the assumption that technological power equals wisdom.

In product development, this shows up as the difference between building features because AI makes them possible versus building features because they serve human flourishing. The stewardship question forces us to ask not just “can we?” but “should we?” and “to what end?”

Practical Implications for Product Builders

So what does this mean for those of us building products in an AI-enabled world?

First, it means getting serious about the wisdom question. Mollick’s co-intelligence framework is helpful, but it needs theological grounding. AI partnership isn’t just about efficiency, it’s about ensuring that our use of AI capability serves love of God and neighbor.

Second, it means designing for human flourishing, not just human preference. AI can predict what users will click on, but it can’t determine whether clicking on that thing actually serves their long-term spiritual formation. That requires human judgment informed by wisdom.

Third, it means accepting that co-intelligence is inherently messy. The Proverbs model of seeking counsel assumes disagreement, iteration, and the need for ongoing discernment. AI partnerships that work will feel more like conversations than commands.

In our recent experiments, the most successful AI implementations have been the ones that generate multiple options rather than single recommendations, that surface uncertainty rather than hiding it, and that make their reasoning transparent so humans can engage with it meaningfully.

The Long View

Mollick is right that we’re entering an era of co-intelligence. But I think the Christian perspective adds something crucial to his framework: the recognition that intelligence without wisdom is dangerous, and wisdom without love is meaningless.

“If I… can fathom all mysteries and all knowledge… but do not have love, I am nothing.” (1 Corinthians 13:2, NIV)

Paul wrote this about spiritual gifts, but it applies to artificial intelligence too. The goal isn’t just more capable AI systems, it’s AI systems that help us love God and neighbor more effectively.

That’s a higher bar than efficiency or even intelligence. But for those of us building products that serve spiritual formation, it’s the only bar that matters.

The co-intelligence era is coming whether we’re ready or not. The question is whether we’ll approach it with the wisdom of Proverbs or the folly of Babel. I’m betting on Proverbs, but only if we’re intentional about what that actually means in practice.

¹ Mollick, Ethan. “Co-Intelligence: Living and Working with AI” (Portfolio, 2024). See also his ongoing research at OneUsefulThing.org.

Photo by Mindfield Biosystems on Unsplash

View all posts