Is Artificial Intelligence Neutral?
The LoCK plugin openly begins from Christian commitments. It does not hide the fact that Jesus Christ and God’s Word govern the way it approaches questions. That makes the objection raised at the end of the previous article understandable: if the LoCK plugin begins with a Christian foundation, does that make it biased while a general-purpose AI chatbot remains neutral?
At first, a general-purpose chatbot can certainly appear more neutral. Ask about a difficult moral or cultural issue and it may present several perspectives, explain the reasoning behind each, and avoid declaring one of them authoritative. Compared with a tool such as the LoCK plugin, which states its governing commitments openly, that can look like the more objective approach.
But neutrality is more difficult than it first appears.
Generative AI does not approach a question from nowhere. It works with human language and human-produced information, while people make decisions about how these systems are trained, instructed, and expected to respond. Different systems handle those decisions differently, but none escapes them entirely.
The presence of assumptions is only part of the issue. We also need to understand what those assumptions are doing.
Neutrality Is Harder Than It Looks
Generative AI models learn from enormous amounts of human-produced material. Books, websites, articles, conversations, and other forms of language help these systems learn patterns that allow them to generate useful responses. Those sources contain facts and observations, but they also contain the disagreements, values, assumptions, and competing beliefs of the people who produced them.
The process does not end with the material used in training. People also make decisions about how an AI system should behave. Depending on the system, those decisions may include instructions, safety boundaries, expectations for helpful behavior, or guidance about how disputed questions should be handled. A chatbot may acknowledge uncertainty, avoid certain kinds of content, or present multiple perspectives when a question is contested.
We do not need to assume that AI developers are secretly attempting to insert a worldview into every answer. The simpler point is enough: generative AI is a human-made technology working with human-produced information and shaped by human decisions.
For some questions, those influences may have little effect. Ask a chatbot to convert inches to centimeters and worldview assumptions will probably have little bearing on the answer. Ask what gives human life value, what makes a marriage good, whether a particular action is moral, or what makes life meaningful, and the situation changes.
Those questions require more than retrieving or organizing information. They involve judgments about what is true, what is good, and what should guide human life.
Presenting several positions without choosing among them does not escape those judgments. In some circumstances, presenting multiple perspectives may be exactly the right and useful thing to do. But deciding which perspectives to include, how to describe them, and whether they should be treated as peers can itself reflect assumptions about fairness, knowledge, and truth.
A response shaped by assumptions is not therefore deceptive or useless. It simply should not be mistaken for a view from nowhere. Usefulness should not be confused with neutrality.
Christians Are Not Called to Stand Nowhere
Once we recognize that generative AI operates with assumptions, we could easily pursue the wrong solution. We might begin looking for an AI system with no assumptions at all, as though faithful discernment requires finding a perfectly neutral place from which every claim can be evaluated.
Christians were never called to such a place.
Jesus Christ was crucified and bodily raised from the dead. The risen Christ declared, “All authority has been given to me in heaven and on earth” (Matthew 28:18). He also said, “I am the way, the truth, and the life. No one comes to the Father except through me” (John 14:6).
Christians cannot set those realities aside whenever we want to examine a question neutrally. They establish who Christ is and why His followers receive His words as authoritative.
The Christian starting point is not neutrality. It is the risen Christ.
Beginning with Christ still allows Christians to listen carefully to people who disagree with us. It also does not make every Christian interpretation automatically correct. Science, history, culture, philosophy, and people who begin from very different beliefs can all contribute observations we should hear carefully.
Christ’s authority should make us willing to examine our own reasoning. We can listen carefully because truth does not threaten Him. What we encounter may reveal a factual mistake, expose weak reasoning, or show that something we have treated as Biblical truth is actually a cultural or personal assumption that needs to be reconsidered.
Such correction does not replace the One who possesses all authority. Christ is not an obstacle that must be removed before Christians can see reality clearly. He is the One under whose authority we seek to understand it.
We do not need to pretend that we have no starting point. We need to remain faithful to the right one.
What Governs the Answer?
Christians can acknowledge the assumptions within generative AI without becoming suspicious of every answer it produces. A chatbot can provide accurate historical information, explain a scientific concept well, or offer a useful observation about human behavior. The presence of assumptions does not make every answer false.
The deeper concern appears when those assumptions begin to shape interpretation.
Consider an AI answer that says a good life is one in which a person becomes the most authentic version of himself. There may be worthwhile observations within that answer about honesty, integrity, or resisting unhealthy social pressure. But the claim still leaves something unresolved: Who determines what the self ought to become?
Or a chatbot might describe morality primarily in terms of reducing harm. Christians can recognize the importance of protecting others from harm while still asking: Is avoiding harm enough to define what is good?
Questions like these help reveal what is governing the answer. What view of human beings is being assumed? How is good being defined? What counts as truth? What source or standard is ultimately allowed to settle the question?
We are not merely looking for evidence that an AI chatbot is biased. We are trying to understand how facts and observations are being interpreted.
The same examination must turn inward. Christians bring assumptions to questions too. Personal preference, political commitments, church traditions, cultural habits, and individual experience can all be mistaken for Biblical truth.
Recognizing assumptions within AI should lead toward humility, not greater confidence in our own. Christian discernment cannot simply replace the assumptions in an AI answer with whatever assumptions we already hold.
Our assumptions need examination too.
For Christians, that examination takes place under the authority of Christ and God’s Word. We should be willing to affirm what is true, correct what is mistaken, reject what contradicts what God has revealed, and keep investigating when the evidence does not justify a confident conclusion.
That is very different from simply asking whether an AI answer agrees with us.
From Neutrality to Discernment
Generative AI does not approach every question without assumptions. Neither do the people who build these systems, the human material from which they learn, or the people who use them. This does not make generative AI uniquely untrustworthy. It places these tools within the same human world in which Christians have always been called to exercise discernment.
The LoCK plugin makes its starting point explicit. It begins from the Christian confession that Jesus Christ is Lord rather than claiming to reason from a view with no commitments. Its explicit foundation does not make every answer produced with the LoCK plugin correct. Its governance can be applied poorly, its information can be wrong, and the Christian using it remains responsible for discernment.
There is no simple choice between an AI with assumptions and one without them. We need to recognize the assumptions shaping an answer, examine what they are doing, and bring them before the authority that should govern them.
For Christians, that authority cannot finally rest on personal preference, cultural consensus, or the output of an AI system. Jesus Christ is risen and reigns. His authority provides the foundation from which Christians examine the assumptions within AI answers and the assumptions we bring with us.
Recognizing those assumptions is only the beginning. Christians still need to know what to do with an answer once they receive it. We need to distinguish a factual error from a deeper interpretive problem, know when something should be affirmed, corrected, rejected, or investigated further, and remain responsible for what we ultimately believe.
That is the work of discernment.