Artificial Intelligence Statement

By Douglas Groothuis, PhD

This essay takes up the ethics and significance of generative AI or large language models related to the humanities and generally excludes other kinds and uses of AI. I am not concerned with AI’s ability to sort vast amounts of technical data related to business or medicine or the military, important though these and other areas are. Nor will I consider catastrophic scenarios of AI takeover of life as we know it. Nevertheless, I hope the analysis and principles I offer might apply broadly.

Human Nature, Skills, and Knowledge

As creatures made in God’s image, we create tools and technologies (Genesis 1:26-27). That is our nature and is not a result of the fall. However, as fallen creatures, we often misuse technologies, even with the best intensions. These technologies can extend our God-given dominion over creation and help compensate for some of the effects of the fall in the material realm. Only the work of Jesus Christ, however, can redeem us from the effects of the fall in our standing with God (John 14:6; 3:16; Romans 5:1-8). However, Christ is Lord over all of life, body and soul, church and culture (Matthew 28:18-20; 1 Corinthians 10:31; Colossians 3:17). Thus, we consider how to glorify God in our engagements of technology.

Creation, Fall, Redemption, and Consummation

Our philosophy of AI for education should be placed within a Christian worldview. One way of elucidating the Christian perspective on life is the well-known model of creation, fall, redemption, and consummation. Although not a theologian, Francis Bacon can get us started in this endeavor by his statement about responding to the fall. He wrote in Novum Organum Scientiarum (1620):

Man by the Fall fell at the same time from his state of innocence and from his dominion over creation. Both of these losses, however, can even in this life be in some parts repaired; the former by religion and faith, the latter by the arts and sciences.

Bacon meant—to put the best biblical face on it—that our state of innocence was lost at the fall and so our redemption could only be wrought through true religion and faith in God’s saving plan. There is no technology, AI or otherwise, can grant us a right standing with God. That comes only by receiving the gift of salvation by faith through the finished work of Jesus Christ. However, some of the physical effects of the fall can me repaired ameliorated through human thought, ingenuity, and industry, what Bacon calls “the arts and sciences.” Pain in childbearing was either added or increased by the fall (Genesis 3), but medical procedures and anesthetic can relieve this health detriment without going against God’s purposes. This principle is widely applied in medicine. Consider something as simple as eyeglasses or heart medication or even a heart transplant.

Because of the fall, the world is abnormal, broken, groaning in travail. We can address this in the wisdom and power of God without fighting God himself or the original good order of creation. Neither can we bring about the restoration and perfect justice of the Eschaton through our efforts. So, we resist all utopian approaches.

Some technologies do not remediate aspects of the fall, but extend our dominion over creation, which is part of our purpose as image-bearers of God. Automobiles and airplanes are in this category. But, given this fallen world, cars and planes crash and can be used for mischief of all kinds. As Larry Crabb said, in a fallen world, “There is something wrong with everything.” Bob Dylan, “Everything’s broken.”

But with AI, we switch our focus to communication technologies. The difference with generative AI is that it produces readable text through human engineering at the design level and prompting at the individual level. AI is in a different category than facilitating communication between people through telephone or email. It not only mediates information; it generates information in remarkable ways.

How might generative AI address maladies resulting from the fall related to communication and knowledge? First, the fall created a scarcity of needed resources. Labor is more difficult and goods more elusive. Task may take more time than they would otherwise. Work becomes more arduous: “by the sweat of their brow.”

AI scans available texts and produces documents. It does so rapidly, far more rapidly than a manual search. But this information has to be vetted, since AI generate “hallucinations” (or false information). Since it lacks consciousness, AI cannot read anything. It can only acquire and associate words given a system of computation. It can only scan words for associations in order to produce texts.

Mutatis mutandis, AI writes nothing; it only produces strings of words that can be read by humans.[i] Writing requires consciousness and agency, both of which AI lacks. AI produces texts of varying quality. But having neither consciousness nor agency, it possesses neither knowledge nor discernment. It cannot test for truth through cognition to derive knowledge. Furthermore, AI has no conscience or wisdom. Thus, it cannot discern who needs what knowledge and when. For example, AI programs have counseled people to commit suicide. Several have done so.

AI may be used to assist or refine cognitive processes in writing but it should not be used at the expense of needed human skill development and performance (research skills, analytical-critical thinking, writing competence) or in ways that sacrifice intellectual integrity. Intellectual integrity involves (1) avoiding plagiarism and (2) adequate vetting of information as knowledge.

First, consider intellectual skill development and performance. AI can “edit” prose to change it, perhaps to improve grammar or to add clarity. But a text thus altered is not the work of a human author, but a collaboration of a human who provides raw material and the AI function that (not who) shapes it. The author is not required to craft the right words or style for writing if AI is employed for that work. In this case, the skills needed for authorship are outsourced to an impersonal technology, although the final product may seem entirely human.

A student may use AI to find sources for a research paper, but have no knowledge of the value of the sources. Even if the sources are trustworthy, this would be a case of  epistemic happenstance for the student, since he or she does not know that the sources are reliable. Let me illustrate. A friend asked me about the work of the French sociologist, Jacques Ellul, who I have been reading for nearly fifty years. I asked the Grok AI to summarize his academic work. The result was largely accurate and produced quickly. However, I was intellectually qualified to authenticate the AI text, given my experience. If an eighteen-year-old student of mine had submitted the same prompt to the same AI, he would have likely gotten the same answer. The information would be identical. However, unlike me, the student would not know the information was truthful, given his lack of intellectual background. Thus, a method for deriving true statements can lack knowledge on the part of the user, thus debasing or even obviating learning.

AI and Human Flourishing Under God: Shalom Value

As a Christian university, we desire that our students, staff, and faculty thrive in godly relationships in which we “speak the truth in love” (Ephesians 4:15). We call this “the beautiful community.” This is part of shalom, or living in peace, truth, and goodness among others in the fear and love and knowledge of God. But community and shalom can be undermined by AI and overindulgence in the online world in general. This happens in at least two ways.

First, AI can substitute an impersonal exchange of data between student and AI agent for the person on person interaction of teacher and student. The most extreme example would be if a student uses AI to write a paper and a teacher uses AI to grade the paper. All human contact is therein eliminated and no education is eliminated in this pointless data exchange scam. Or an AI could create a quiz that the student takes and receives a grade. Perhaps the quiz is apt for the class. However, the teacher is not personally involved with the student. He or she may not know anything beyond the raw grade the student receives. This is not authentic teaching and learning, which is discipleship or mentoring.

Second, AI can obstruct relationships between students through the overuse of screens. Instead of meeting together for a study session or a prayer meeting or a playing a board game or frisbee, students may text each other or play video games. Some schools even support video game teams as “sports,” thus completely removing bodily action and physical proximity. Humans need unmediated time together, especially younger people learning social skills pertaining to being Christian students, friends, church goers, and citizens.

Third, AI can accelerate or more make efficient some learning without undermining skill or knowledge in research, writing, and teaching. For these uses, questions should be asked. Is using AI counterfeiting knowledge or contributing to it? For example, asking AI to summarize Kant’s categorical imperative sidesteps a deep engagement of Kant’s writing or even a summary of it in a reputable secondary source. But perhaps an AI could tell you where you could read Kant’s verbatim text online, although the physical book makes for better engagement.[ii] A student might ask an AI to summarize the work of Francis Schaeffer or C. S. Lewis and to list their books. But then the student should ask a teacher or mentor if the report was accurate, since he or she is unlikely to be able verify that. Consider the previous example about Jacques Elull’s work.

Perhaps AI could help a student in developing basic foreign language skills through language drills on vocabulary grammar. However, the teacher would need to verify that the exercises were appropriate and would need to grade the students’ work for substantiative assignments. The key question is: Does AI enhance learning—or does it outsource skills and knowledge, thus robbing potential learners of learning?

Christians should neither adopt new technologies too quickly (and be impatient) or be obstructionists about their wise use (and be cowardly). A thoughtful analysis of how AI contributes to or hinders education must be pondered according to a Christian worldview and by the facts on the ground, which are rapidly changing concerning AI. As usual, we need Spirit-led wisdom from above (James 1:17).

[i] See Mark Warner, More Than Words.

[ii] See Douglas Groothuis, “The Book, the Screen, and the Soul,” The Soul in Cyberspace (Baker, 1997).