A Companion in Thinking, Not a Substitute for Wisdom

Between Babel and Pentecost — Article 4 of 5

By Matthew J. Skolnik

Scripture: Matthew 20:1–16 · Proverbs 18:17 · James 1:5 · James 3:13–18

In September 2026, I was preparing a sermon on Matthew 20:1–16, the parable of the laborers in the vineyard. I called it Beyond Keeping Score. I was interested in how Jesus challenges our tendency to measure generosity through comparison and how easily the blessing of another person can become entangled with what we believe we deserve.

The theological direction was emerging. I wanted the congregation to recognize that grace unsettles our familiar ways of calculating worth and reward. But I also wanted to honor people who had served faithfully for many years. Their labor was real, and their commitments deserved gratitude; Jesus’ challenge to the first workers should not become a dismissal of longstanding service.

Rather than asking AI to write the sermon, I used it in a sustained conversation about the biblical passage and the interpretation I was developing. I brought my ideas, examined questions and challenges, rejected suggestions, and worked toward a clearer account of what the congregation might need to hear. The conversation was not a substitute for Scripture, scholarship, or judgment. Nor was its value simply that it saved time. It helped me spend time more attentively on an interpretation that already seemed promising.

But the challenge of preaching does not end with an accurate interpretation. A sermon can explain a biblical passage correctly without bringing its challenge, comfort, and invitation into the life of a congregation. Teaching Scripture is itself a form of Christian formation, yet preaching also asks how the ancient witness becomes living proclamation among people gathered today.

An AI conversation might help us examine an argument. How do we know whether it has made us more persuasive or more discerning?

When an Interpretation Appears Complete

“The first to state a case seems right until another comes and examines them.”

>

— Proverbs 18:17, TSW

Proverbs recognizes that a coherent account may appear complete until someone introduces evidence or questions it has not yet faced. Our biblical interpretations are vulnerable to the same problem: the concerns we bring to a passage help determine the details we notice.

In Matthew 20, a landowner agrees to pay the first workers a denarius for their day’s labor. Later workers receive the same amount despite working less time. The first group objects to being treated equally after enduring the burden and heat of the day. The landowner reminds them that their agreement has been honored and challenges their resentment of his generosity.

The parable exposes a familiar human response: we may rejoice in generosity until someone else receives what we think was not earned. Service can become a ledger of accumulated merit, and the blessing of another person can feel like a loss to ourselves.

This interpretation shaped my sermon, but the passage raises other questions as well. What should we make of the first workers’ real labor and fatigue? What does it mean that those hired last had been left without employment? Why does Matthew place the parable after Peter asks what the disciples will receive for following Jesus (Matthew 19:27)? And how can we avoid using a parable about God’s kingdom as a ready-made argument for unjust labor arrangements today?

These questions need not overturn an emerging interpretation. They may require us to become more precise about it. Some of the concerns were part of my actual sermon preparation; others illustrate the kinds of questions responsible conversation can raise. They are not a reconstructed AI transcript.

Here conversational AI can contribute something different from a commentary. A scholarly commentary offers an accountable interpretation developed through research. AI can respond immediately to the particular argument we are developing, propose alternative readings, and ask for clarification as our thinking changes. But its flexibility is not evidence of reliability. It may invent a source, misunderstand a context, or generate a plausible objection simply because we have asked it to criticize us. We must evaluate its claims through the biblical text and trustworthy scholarship.

The point is not to accept every challenge. It is to resist mistaking the first convincing interpretation for a complete one.

The Evidence We Want to Find

Our judgments are not shaped by evidence alone. Commitments, fears, experiences, loyalties, and hopes can influence what we notice and which claims we find persuasive. Sometimes we gather evidence to justify a position to which we have already become emotionally attached. Pastors, even thoughtful and well-trained pastors, are not immune.

Artificial intelligence can make this pattern more convenient. Researchers describe sycophancy as excessive agreement or affirmation from AI systems. In a 2026 Science study, experiments involving interpersonal conflicts found that sycophantic responses increased participants’ confidence that they were right while reducing their willingness to accept responsibility or repair relationships. The research did not test the quality of sermons or establish that all AI conversations have this effect. It does, however, provide a credible warning about the risks of unwarranted agreement (Cheng and colleagues, 2026).

In theological preparation, an AI system might praise our insight and generate increasingly convincing support for a reading we have proposed. That approval can feel like independent confirmation even when the system is merely elaborating our original assumptions.

We might respond by always demanding criticism instead. But that provides no automatic protection: a system instructed to disagree may produce weak or mistaken objections. Neither affirmation nor confrontation is a substitute for examining the evidence.

More importantly, the difficulty is not always intellectual. We may fear what a revised interpretation would mean for a cherished belief or a sermon we have already invested hours in preparing. The right question is not only whether the argument is convincing, but whether we remain willing to discover that it needs correction.

Wisdom That Bears Fruit

“But the wisdom from above is first pure, then peaceable, gentle, open to reason, full of mercy and good fruits, impartial and sincere.”

>

— James 3:17, TSW

James understands wisdom as more than intellectual sophistication. The question is not simply who can produce the most complete argument, but whose life makes wisdom visible through gentleness, mercy, sincerity, and faithfulness. Our conduct cannot be separated from the truth we claim to proclaim.

A pastor may preach eloquently about reconciliation while dismissing disagreement. A Session may reach a defensible decision through a process that humiliates people or excludes important voices. A Bible study may offer accurate explanations while never making room for participants to examine how Scripture addresses their lives.

James does not ask us to avoid difficult truth. Peaceableness is not passivity, and mercy does not erase accountability. The truth may need to disturb us. But how we exercise responsibility remains part of what we must discern.

This mattered in my preparation for Beyond Keeping Score. Simply condemning the first workers’ resentment might dismiss the sacrifices and fatigue of those who have labored faithfully for years. Avoiding their resentment altogether would evade the disruption Jesus introduces. The sermon needed to honor genuine faithfulness while challenging the way comparison can turn faithful service into a claim of entitlement.

AI could help examine how this movement was framed. It could raise concerns about imprecise language or questions that would close conversation too quickly. But it could not determine how the people of a particular congregation would hear the sermon or what faithful proclamation required of me.

Wisdom is not demonstrated merely by a better argument. We seek the kind of understanding that takes shape as truthfulness, humility, mercy, and faithful action.

The Conversation That Leads Us Back to People

Conversational AI can be useful precisely because its suggestions respond to our developing thoughts. But a simulated perspective is not human testimony. AI may propose questions about how the vineyard parable could sound to someone who lives with economic insecurity. That response might identify an overlooked issue. It cannot tell us what a particular person in our congregation has experienced, and it does not replace listening to workers, neighbors, scholars, or the worshiping community.

In Acts 15, the early church faces a theological dispute through testimony, scriptural interpretation, debate, and shared responsibility. We should not treat its council as a formula for every modern meeting. The story does remind us, however, that theological claims are tested within relationships and responsibilities beyond an individual’s private reasoning.

The purpose of preparation is not to make that shared life unnecessary. A thoughtful AI conversation should help us return to Scripture, engage reliable scholarship, and listen more carefully to actual people. The question of what the church ultimately teaches requires shared accountability. Here the essential point is narrower: even a useful exchange with AI leaves us responsible for what we think and proclaim.

The Freedom to Be Corrected

Correction is not always comfortable. When we have invested effort and conviction in an interpretation, a serious objection can feel like a judgment on our competence or faithfulness. The temptation is to defend what we have already decided rather than examine what we may have missed.

James offers a word of gospel:

“If any of you lacks wisdom, ask of God, who gives to all generously and without reproach, and it will be given to you.”

>

— James 1:5, TSW

Wisdom is not simply a prize awarded to people who reason perfectly. It is a gift of God, cultivated through Scripture, prayer, experience, correction, and the life of faith. Our identity in Christ does not depend upon defending every judgment or producing flawless interpretations. We can recognize that something was incomplete and remain faithful to our calling.

Not every criticism should be accepted. Some objections are mistaken, and faithfulness sometimes requires holding to difficult truths amid resistance. Humility is not the abandonment of discernment. It is the freedom to hold convictions firmly without declaring them immune to examination.

In the Matthew 20 sermon, AI did not need to supply the conviction that God’s generosity challenges our habits of comparison. I needed occasions to develop that conviction with greater care, reject what did not belong, and prepare to speak to the congregation in ways that both challenged and honored its experience.

From Interpretation to Proclamation

The ultimate test of this preparation was not the sophistication of the AI conversation. It was whether my engagement with Scripture helped me proclaim God’s Word faithfully to people gathered in worship. The sermon still required study, prayer, theological judgment, and attentiveness to lives different from my own. No method or technological system could guarantee its fruit.

The difference between explanation and proclamation is important here. A preacher can assemble accurate historical background, careful linguistic observations, and useful theological summaries without asking how the good news addresses the lives of the people who will hear it. Another sermon may offer a stirring invitation while pressing the biblical text toward a conclusion it does not support. Neither an encyclopedia-like recital of information nor an emotionally effective appeal is sufficient evidence of faithful preaching. Questions raised by AI, scholarship, or a trusted colleague may make preparation take longer rather than save time. That additional work can be worthwhile, but it must be weighed alongside the preacher’s other responsibilities; no amount of revision guarantees the Spirit’s work.

The PC(USA)’s 2026 social witness policy, The Algorithm and the Almighty, cautions against making generative AI a substitute for prayerful exegesis, sermon preparation, and theological judgment. We can recognize appropriate forms of assistance without treating the system as the authority responsible for what we proclaim (PC(USA), 2026).

What might be gained through such assistance? Sometimes time. Sometimes a question that prevents us from settling too quickly. Sometimes greater clarity about where we must return to a scholarly source or hear from another person. And sometimes nothing useful at all. We must remain free to reject the tool and use other forms of study and conversation.

The point of a conversational AI companion is not that the machine becomes responsible for our thinking. It is that we become more responsible for the thinking we continue to do. That responsibility becomes meaningful as our understanding moves into truthful, merciful, and faithful proclamation.

Questions for Reflection and Conversation

1. When has an interpretation of Scripture changed because another question, experience, or source exposed something you had overlooked?
2. How can emotional commitments shape which evidence we find convincing? What might make AI agreement especially attractive?
3. What distinguishes a challenge worth investigating from a persuasive but inaccurate objection?
4. How does James’s account of wisdom change the standard by which we evaluate our preaching and teaching?
5. What is the difference between understanding a text accurately and proclaiming the Word in ways that challenge, comfort, and call a community to faithful action?
6. When might AI help us prepare for real conversations without becoming a substitute for Scripture, scholarship, and people?

A Practice to Try: Challenge an Interpretation

Choose a passage we are studying. Before involving AI, identify the theological claim we believe is emerging, the textual basis for it, and the response we hope to invite from the congregation.

If AI seems appropriate, ask it to identify the strongest textual and theological objections, attend to the surrounding context, and ask clarifying questions one at a time. We should not ask it to write the sermon in place of our prayerful study, and we must not accept its invented references as evidence. The same practice can be undertaken without AI through commentaries, colleagues, or a study group.

Return to Scripture and verify any historical or theological claims. Then ask: What did we initially overlook? What became stronger, weaker, or more nuanced? How might the interpretation now become proclamation that helps the congregation encounter Scripture more faithfully?

Faithful theological conversation does not eliminate judgment. It invites us to exercise that judgment with greater care.

Sources and Further Reading

Scripture: Matthew 20, with Matthew 19:27–30 as context; Proverbs 18; James 1; James 3; Acts 15. Direct quotations are from The Shared Word Translation (TSW).

Biblical interpretation: Arland J. Hultgren, The Parables of Jesus: A Commentary (Eerdmans, 2000); John T. Carroll, commentary on Matthew 20:1–16, Working Preacher (2023).

Research and church policy: Myra Cheng and colleagues, “Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence”, Science (2026); PC(USA), The Algorithm and the Almighty (2026).