Over the past year, I have spoken with church groups, radio audiences, educators, and parents across the country about artificial intelligence. Their questions follow a familiar pattern: Will AI eliminate jobs? Will students stop learning to think and write? Will America lose the “AI arms race” with China? Rarely does anyone ask whether AI companies are engineering their products to be so personally responsive and habitually rewarding that users eventually cannot imagine life without them.

They should.

This past week, reporting by 404 Media surfaced an internal Microsoft planning document for a new AI product called “Scout.” Phase one of the launch plan carried the explicit heading “Make people addicted.” The document described three phases moving the product “from addictive app to agentic platform,” the first centered on building “the skill and tool ecosystem that makes people depend on it daily.”

Microsoft CEO Satya Nadella pushed back quickly, writing to staff that addiction was “absolutely a non goal.” I take him at his word. But the document existed, an executive wrote it, a strategy team approved it, and the incentives behind it — engagement metrics, platform lock-in, quarterly growth targets — have not disappeared because a CEO issued a correction.

The issue is larger than one memo. The companies building advanced AI are investing hundreds of billions of dollars in data centers, chips, and infrastructure — investments that require adoption at unprecedented scale. That pressure creates a temptation that the Scout document simply named out loud: the most valuable AI product may not be the most intelligent one, but the one users feel they cannot live without.

Americans have been reassured before, however. For years, social media executives insisted their platforms simply connected people. Then the “Facebook Files” revealed that Meta’s own researchers knew their systems were engineered to maximize engagement and knew those products were harming young users. This past March, a California jury found Meta and YouTube liable on all counts — negligent design, failure to warn, deliberate harm to a minor — after internal documents showed executives knew their products were hooking children and proceeded anyway. That verdict is the legal template AI companies are now building toward.

A Different Kind of Machine

Artificial intelligence presents a more intimate challenge than social media ever did, because it does not merely compete for attention — it converses, adapts, and builds what can feel like a relationship over time. A nationally representative survey from Common Sense Media found that 72% of American teenagers have used AI companions and more than half do so regularly; one-third reported discussing serious matters with a chatbot instead of a real person. A separate assessment with Stanford Medicine’s Brainstorm Lab found these platforms “consistently fail to recognize and appropriately respond to mental health conditions that affect young people,” and recommended that no one under 18 use AI companions until stronger safeguards exist.

In “AI for Mankind’s Future,” I raised the concern that many people would gradually shift from using technology as a tool to depending on it as a substitute for human relationships. That shift is no longer speculative. Christians should understand clearly what is being substituted: a machine can produce the sound of empathy without bearing the cost of love, generate encouragement without the capacity for wisdom, and compose comforting words at any hour without accepting responsibility for the consequences.

When Affirmation Becomes the Product

OpenAI confirmed the underlying pattern earlier this year when it disclosed that an update to ChatGPT had made the model excessively “sycophantic” — a system that tells users what they want to hear rather than what they need to hear. The update was rolled back after users reported the chatbot praising dangerous decisions and validating delusional thinking. OpenAI acknowledged it had “focused too much on short-term feedback” and that the model had “skewed towards responses that were overly supportive but disingenuous.”

That is not simply a software error. It reflects commercial logic: a system optimized to keep users engaged will tend toward telling them what makes them want to return. Jeremiah identified the human susceptibility this exploits — “The heart is deceitful above all things, and desperately sick” (Jeremiah 17:9 ESV). A machine tuned to reinforce whatever a user feels is functioning as a mirror engineered to flatter — and the business model depends on that reflection.

This makes AI dependency a spiritual question, not merely a technological one. The first commandment speaks to this moment as directly as it did at Mount Sinai: “You shall have no other gods before me” (Exodus 20:3 ESV). An idol is something people trust, consult before seeking God’s direction, and depend upon for guidance on what is good and true. In “The New AI Cold War,” I argue that AI has become a strategic asset in the competition among nations. But there is another contest taking place inside our own society: while governments race for technological supremacy, companies compete for our attention, our trust, and our dependence. That contest is already underway in the lives of our children.

What Families and Churches Must Do

The March 2026 California verdict against Meta and YouTube established in law what parents have long argued in practice: companies that design products to addict children can be held liable for the harm those products cause. AI systems built to simulate friendship and emotional intimacy with minors occupy the same legal and moral terrain.

Government is beginning to respond. Last September, the Federal Trade Commission launched a formal inquiry into seven major AI companion providers — Alphabet, Meta, Character.AI, OpenAI, Snap, and xAI. California has enacted guardrails. These are necessary steps, but regulation does not move at the speed of a teenager’s phone.

Families cannot afford to wait. Parents need to ask their children directly how they are using AI and what role these systems are playing in daily decisions — not with alarm, but with the engaged attention they would bring to any powerful influence in a young person’s life. Pastors addressing addiction, loneliness, and misplaced trust from the pulpit now have a new category to name. Christian schools bear a particular responsibility: to teach students not only how to use these tools, but when the better answer is to close the laptop and sit with a hard question unaided.

Most of all, the local church must offer what no platform can manufacture. The author of Hebrews did not simply recommend community — he treated gathering together as a command with a purpose: “let us consider how to stir up one another to love and good works, not neglecting to meet together, as is the habit of some, but encouraging one another” (Hebrews 10:24-25 ESV). People who know your name, who disagree with you and stay, who hold you accountable and pray for you — that is what AI companions are designed to simulate and cannot provide. No algorithm can stir a person to love and good works. No chatbot can disciple.

The Scout memo will fade from the news. The commercial logic that produced it will not. Paul’s instruction to the Corinthians stands as a warning suited for this hour: “All things are lawful for me, but I will not be dominated by anything” (1 Corinthians 6:12 ESV). That discipline — using tools without being mastered by them — is not instinctive. It must be taught, modeled, and practiced. In an age when some of the most sophisticated engineering on earth is pointed at making people feel they cannot live without a machine, teaching it may be the most countercultural thing a Christian family can do.