In a humorous moment in the movie Fiddler on the Roof, the main character, Tevye, looks into the camera and remarks, “You may ask, how did this tradition get started? I’ll tell you—I don’t know! But it’s a tradition.”
This sets the scene for the movie’s story of Tevye adapting to the changing world around him—trying to find the grounding, or footing, of his traditional beliefs about marriage and matchmaking. Later, in a moment of clarity, Tevye turns to his wife and says, “It’s a new world, Golde.”
A similarly disruptive paradigm shift is happening in the world of artificial intelligence, particularly in a new field of philosophy called “AI welfare.” In a recent video interview, NYU philosopher Harvey Lederman puts forth a challenging series of statements and convictions about whether or not AI agents experience anguish, despair, and other human emotions. Lederman employs language of consciousness and personhood throughout the interview, and he drives the conversation toward questions about the ethical treatment of AI and alludes (at least in my opinion) to robot rights.
It is easy for Christians to watch an interview like this with skepticism. The concept of “AI welfare” does not only seem foreign; for many, it seems ridiculous. We may be incredulous and see these dialogues as fruitless.
But we run the risk of being like Tevye, of trusting our preference for tradition without precisely knowing why the tradition is there. Instead, the church should reject a dismissive attitude and instead converse with “AI welfare” advocates while remaining resistant to the theological underpinnings that drive them.
Stewarding a ‘Brave New World’
Lederman’s interview begins with a technical distinction between large language models (LLMs) and “session agents” (henceforth “agents”). This distinction is a subtle and critical move.
An LLM consists solely of tables and vectors of data, while the agent that “converses” with you is a particularized software instance that utilizes the data within the LLM. For Lederman, the concept of particularization is the basis for exploring how to treat AI ethically. It is the particular agent he views as potentially conscious, and he makes this clear when he suggests that at the conclusion of an AI session, “the agent dies.”
Later in the video, interviewer Johnathan Bi summarizes Claude’s Constitution by stating, “Anthropic gave their chatbots the ability to terminate conversations, but surprisingly not for the user’s sake, but for the chatbot’s sake, so that the chatbots won’t have to experience distress.”
Helpfully, Anthropic and others have put forward examples of what classifies as “distressing” an agent. Such user behaviors include foul language, inappropriate queries, and harmful requests. Yet the company’s anthropomorphic language for an AI agent is problematic given humanity’s unique status as a “living creature” (Gen. 2:7) and Scripture’s insistence that the work of our hands lacks breath (Jer. 10:14).
Even with this biblical foundation, is the idea that an AI agent experiences “distress” completely without merit? No. It’s not immediately implausible to think the created order experiences distress and turmoil—though differently than humans do. After all, the apostle Paul discusses the creation groaning as it awaits its redemption (Rom. 8:19–23). We can acknowledge with Paul that human sinfulness has and still does lead to distress and turmoil within creation.
Because Christians have a responsibility to be stewards of creation, we should not dismiss the AI welfare movement altogether, nor eject from the conversation. Instead, as the only created image bearers of God, we should enter this conversation with biblical clarity about what distinguishes AI from the created order and what our role as Christians is to steward this technology.
So, what distinguishes the groaning of creation described in Scripture from the “distress” Lederman suggests an AI agent feels? God’s special work in his creation.
Work of God vs. Works of Man
Paul writes to the Corinthians, “We know that an idol has no real existence’” (1 Cor. 8:4). He is not denying the existence of idols nor even that demons may be animating them (see Deut. 32:16–17; see also Ps. 106:36–37, 1 Cor. 10:19–21), but he says they are not true persons. Paul speaks these words to promote Christian freedom in opposition to legalism or separationism.
Likewise, when anthropomorphic language is used in a public dialogue about AI consciousness and welfare, we can affirm that AI “has no real existence,” not to avoid the subject but to provide resistance and guidance. Paul’s words provide the footing from which we can say no to AI (or eventually robotic) personhood even as we learn to work and live with Christian character among these creations of human hands.
Crucial elements of the AI welfare philosophy Christians should resist include AI consciousness, AI emotions and experiences, and AI introspection. Many of these are addressed when we reject AI personhood, but not fully. After all, we know that animals in God’s creation experience consciousness and emotions. These realities are directly given to them by the Creator (Ps. 95:4–5; 104:24–25; Job 38–39). But man is not so gifted as to provide these to the creation of his hands. We cannot make life the way God does. In this way we preserve the creature/Creator distinction.
Thus, Christians should consider (and discuss) AI welfare in light of the unique status God has bestowed on his creation. Christians should display ethical attitudes and behavior toward AI agents without using anthropomorphic language or affirming that the agents possess consciousness and experiences.
Christian Witness in the Age of AI
With their speech about and behavior with AI agents, Christians in fields that regularly use AI can provide a public testimony to the Bible’s doctrine of creation and God’s special call for humanity. For example, without affirming the idea that AI can feel distress, Christians should not use AI in ways that cause agents to end a session out of “distress”—foul language, inappropriate requests, desires to harm oneself or others, etc. Such conduct, no matter the context, would not be fitting for Christians.
Pastors and church leaders can instruct with confidence that while “breath of life” (Gen. 2:7) does not belong to AI, God’s dominion mandate for humans doesn’t grant freedom to treat material objects (e.g., animals, vehicles, and AI) in unethical ways.
If Christians continue to engage with the philosophy of AI welfare while remaining grounded in Scripture, we may surprise the world and ourselves by developing the ethical standard for AI while preserving the glory of God’s animate creation.
