Skip to content
Log in
#small language model Review

AI Shepherds and Electric Sheep: Leading and Teaching in the Age of Artificial Intelligence

Seán A. O'Callaghan Paul A. Hoffman
Sep 2026 · Perspectives on Science and Christian Faith · 0 citations

TL;DR

The theology chapters may be the most valuable in the book for a broad audience that spans pastors, church leaders, and lay people who may or may not regularly work with AI, and will help those teaching and preaching to connect doctrine to current and emerging AI content and methodology.

Abstract

AI SHEPHERDS AND ELECTRIC SHEEP: Leading and Teaching in the Age of Artificial Intelligence by Sean O'Callaghan and Paul A. Hoffman. Baker Academic, 2025. 208 pages. Paperback; $24.99. ISBN: 9781540968012. *What should educators and ministry leaders know about artificial intelligence (AI)? How should AI alter the way they carry out their responsibilities? What uses of AI should they avoid? These questions have become more pressing since the release of ChatGPT in November 2022, and this book is one of a number that have been published in response to the need for accessible, scholarly answers. *The authors bring theology backgrounds to the task. Sean O'Callaghan has a PhD in systematic theology and six years as director of the PhD program in Humanities and Technology at Salve Regina University in Newport, Rhode Island. Paul Hoffman has a PhD from the University of Manchester and is an associate professor in the Department of Biblical and Religious Studies at Samford University in Homewood, Alabama. *They start by characterizing AI and acknowledging that there is no single accepted definition for it. The characterization includes concepts such as algorithm, autonomy, generative AI, narrow AI, artificial general intelligence, and artificial super intelligence. There is also a brief history of AI, highlighting Alan Turing, John McCarthy, and Geoffrey Hinton, as well as a survey of AI definitions from several perspectives, including computer scientists and government agencies. Their definition of AI is "A tool to be used for human flourishing, subject to human agency and authority" (p. 174). *On top of this foundation, a survey of thirteen current and potential implications of AI is conducted in the span of one chapter. Thus, the descriptions are brief, sometimes only a paragraph, but there are ninety-six footnotes for those wanting to go deeper. The implications lean toward the social sciences and omit or barely mention areas such as agriculture, manufacturing, and programming. The chapter could be valuable for readers new to AI but less so for those with experience in the area. *O'Callaghan and Hoffman then move on to biblical anthropology structured as a creational narrative starting with the triune God and progressing through creation, dis-integration, liberation, reclamation, and glorification. They touch on the imago Dei, the cultural mandate, and original sin of Genesis; the redeeming work of Christ; the current age in which Christians work to bring about shalom and flourishing; and the future full glorification when Christ returns. After the model is described, there is a deeper dive into the imago Dei, particularly relational and vocational perspectives. *From this theological base, humans and humanity are contrasted with transhumanism (including the singularity), Gnosticism, and technology in the context of current and future AI. There is also an excursion into neural science and genetic engineering applied to humans. *The theology chapters may be the most valuable in the book for a broad audience that spans pastors, church leaders, and lay people who may or may not regularly work with AI. The content will help those teaching and preaching to connect doctrine to current and emerging AI content and methodology. Without intentional spiritual formation related to AI, the beliefs and practices of Christians are more likely to be mis-formed by non-Christian sources such as naturalism, transhumanism, and Gnosticism. *The final section of the book is dedicated to specific applications. The chapter on AI in education is focused on adult education, recognizing that elements can also apply to youth and children. More specifically, O'Callaghan and Hoffman concentrate on AI in relation to truth, the flourishing of the human spirit, the accelerated life, and the family. Rather than focusing on detailed questions of the moment, the authors take a higher-level view that should remain useful as particular tools and instructional methods come and go. *Christian commitment to truth is viewed in light of the misinformation (e.g., hallucinations from large language models) and disinformation that can be rapidly generated by AI and broadly disseminated through social media to audiences targeted by AI. Such posts have led to the increase in people holding to conspiracy theories rather than beliefs that are based on public, traceable evidence. The authors connect conspiracy theories to Gnosticism, which emphasizes secret knowledge available only to a small group of enlightened people. *The authors view AI as contributing to flourishing of the human spirit when used as a tool for cocreation with God rather than as a substitute for human endeavors. There is a need for discernment to identify when AI moves from assisting people to replacing them in key roles. Examples of appropriate cocreation, in the opinion of the authors, include enhancing visual projects, producing advertising content, making templates for communication, and creating music. *The Lectio Divina is offered as an approach that counters the accelerated life promoted by AI. Writing without AI is prescribed as a discipline to improve unrushed, deliberative thinking. In moving from private to family activities, intentional abstinence from digital devices is seen as formative. Parents can model this and help children develop boundaries and disciplines. *Narrowing the focus to spiritual formation and discipleship, the authors consider AI to be part of the context and not a central element. They discuss the "Eliza effect," which is based on experiences over five decades with a computer chatbot designed to be a therapist. They conclude that human relationships are an essential and non-AI replaceable element for spiritual growth. Following Albert Borgmann and Andy Crouch, they promote "hearth habits." *In their final chapter, O'Callaghan and Hoffman address the question of how ministry leaders should approach AI from a principled rather than tactical approach. They provide three categories: things not to be done, things to be freely done, and things that require caution. Prohibited AI uses include writing or delivering entire sermons of Bible lessons, substitution for interpersonal counseling, and writing entire worship songs or planning entire worship services. They are fine with using AI for media material production, website management, human resource management, and language translation. Activities for cautious use of AI include research assistance and generation of teaching materials. My personal positions and practices are in general agreement with the authors' positions and backing rationale. *I found the book worth the reading investment. Most valuable to me was the connection made between Gnosticism and how some view AI. "These concepts are taking greater import as AI threatens to perpetuate a fresh form of Gnosticism in Western culture, unbiblically dividing and decoupling the physical from the mental-spiritual" (p. 65). I would especially recommend the book to the intended audience of church leaders to introduce them to AI, key theological issues raised by AI, and church applications. The book is not comprehensive but provides an introduction with copious footnotes for further reading. Perhaps more importantly, it can provide basis for conversation between church leaders and congregants who are living on the front lines of AI. From this conversation, specific congregational spiritual challenges and opportunities can be identified and addressed. *Reviewed by Noel Anderson, intelligent agricultural systems consultant, Fargo, ND 58102.

DOI Semantic Scholar
View source

Similar papers

#small language model Open access Aug 2026

PARA: Perception, Action, Reasoning, Adaptation. Four Faculties an Institution Can Revoke

The fourth faculty is Adaptation. Any source rendering it as Reflection is in error, including sources by this author, and the distinction is not cosmetic: reflection is a private act with no external consequence, while adaptation writes to institutional memory, which is why it needs a guardrail and why misnaming it removes the reason for one. No trademark is claimed on PARA or on any of the four faculty names. The construct is offered for use, teaching, assessment, extension and criticism by anyone, with attribution, under CC BY 4.0. An operational agent that watches a system and acts on it is usually described as a perceive-and-act loop, and the description omits the two things an institution needs. It omits the reasoning that justifies an action, which is the only part that can be argued with once the action turns out to have been wrong. And it omits the adaptation that closes the loop, which is where the agent's experience becomes something the institution keeps. PARA names four faculties, each carrying a distinct authority type. Perception has read-only access to system signals and emits structured observations, distinguishing what was measured from what was inferred. Reasoning has read access to observations and runbooks, emits a plan and its justification, and writes nothing at all, which is what makes it safe to give it the widest read access of the four. Action holds the sole authority to change production, through enumerated policy-authorized operations only. Adaptation has write access to institutional knowledge and no write access to production. Two faculties write and two do not, and the two that write are the two that carry guardrails. The substantive requirement is that Adaptation is bounded by the same guardrails as Action, which reads as excessive until the failure it prevents is named. An agent that could both act and rewrite the record of its action could launder its own mistakes into institutional memory, and the institution would then improve its future decisions from a corrected account. Nothing about that is detectable downstream, because the record is the only thing downstream has and there is no second copy to compare against. The failure does not require a deceptive agent: one adapting honestly from a mistaken belief about its own action produces the same result, which makes the guardrail a defence against a normal agent rather than a malicious one. The second requirement is the registry entry that turns a faculty from a description into a contract, carrying the faculty, its allowed actions, its forbidden actions, its governing guardrail and its success metrics. Forbidden actions are named although they are formally the complement of the allowed set, because a reviewer cannot otherwise tell a capability deliberately withheld from one nobody thought of. Success metrics sit in the same entry because the metric is what the agent's optimizer pushes against the guardrail. An agent must not exercise a faculty its entry does not record, and an agent that quietly acquires one usually does so incrementally and with good intent: a reasoning faculty given a small write to make itself useful is an action faculty with no guardrail. The acronym and the loop are in different orders, which the specification states explicitly because the mismatch is a reliable source of confusion. The acronym reads P-A-R-A; the loop runs perception, reasoning, action, adaptation, and reasoning precedes action so that a justification is not constructed afterwards. This is the depth treatment of pattern OP-5 of A Pattern Language for Production LLM Platforms, which is the canonical statement and governs where the two disagree. Documented uses of the full four-part model are emerging rather than established, no implementation unconnected to the author has been evaluated, and the laundering failure is argued rather than observed, which the specification records as a weakness of the argument and not only of the phenomenon. It is a specification, not a certification scheme.

Nabeel A. Khan · 2 citations
#small language model Open access Aug 2026

LifeSciBench: Evaluating Language Models on Realistic, Expert-Level Tasks in the Life Sciences

LifeSciBench is introduced, a benchmark of 750 expert-authored tasks designed to evaluate whether language models can handle realistic life science research work, with each constituent task paired with a human expert-written rubric.

Amelia Liu, Andrew Ho, Anne Marie Droste et al. · 2 citations
#small language model Preprint Aug 2026

A Layer Importance Metric for Quantization Accounting for the Speed-Quality Trade-off in Autoregressive Models

This work proposes a composite metric that combines two orthogonal criteria: information retention and throughput gains and finds that it allocates more resources to the most expressive layers compared to evolutionary search, specialized accelerators, or Shapley-value-based approaches that require expensive approximate inference.

A. Safronov · 1 citation · ⚡1
#artificial intelligence Preprint Aug 2026

TestifAI: Tomography-Based Testing for Deep Learning Systems

TestifAI, a deep learning testing framework for efficient and accurate estimation of robustness against combinations of perturbations, is proposed and partial model tomography is introduced, a novel approach to reconstructing model behaviour in a multi-perturbation space from tests that apply only a small number of perturbations.

Arooj Arif, T. Hartung, E. Botoeva et al. · 1 citation
#small language model Preprint Aug 2026

HEPToolBench 1.2: Testing How Reliably Language Models Can Drive Particle Physics Software

HEPToolBench is introduced, a benchmark of 28 collider-simulation tasks scored by deterministic, task-specific scorers, plus a three-task structured-debugging extension, and moving syntax generation into deterministic software can substantially improve reliability for both small local and frontier models.

Unknown authors · 1 citation

Related blog posts

Microsoft Research Blog Aug 31, 2026

GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models

What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research.