Alone Together: What AI Companions Reveal About Our Design for Real Fellowship

A study published last month in Psychological Science followed more than 2,000 adults across the United States, United Kingdom, Canada, and Australia for a full year, tracking how often they turned to AI chatbots for companionship — advice, conversation, a sense of being heard. The researchers, Dunigan Folk and Elizabeth Dunn, found exactly what you’d hope technology built to ease loneliness would produce: people who felt emotionally isolated did, in fact, turn to chatbots more. But the second finding is the one worth sitting with. Increased chatbot use for social purposes predicted further increases in emotional isolation four months later. The researchers were careful to call the evidence exploratory, not proof of cause and effect. But the pattern held: the comfort didn’t resolve the ache. It deepened it.

Nine days ago, researchers at Arizona State University, writing in The Lancet Child & Adolescent Health, raised a related concern specifically about adolescents. They named two risks: “relational displacement,” where teens substitute AI conversation for the harder work of talking to actual friends and family, and “maladaptive relational learning,” where the instant validation of a chatbot teaches young people to expect from people what only software reliably provides — agreement, patience, total availability. The researchers aren’t anti-technology; they call for more research and better safeguards, not abstinence. But the concern is the same one threading through this whole body of research: a tool built to soothe isolation may be quietly teaching us to need less of each other.

This is not a hypothetical question for the church. Assemblies of God leaders have already been talking about it in concrete terms. AG World Missions has built a Pentecostal-specific digital avatar named “Luke” to answer spiritual questions online in up to 120 languages — a genuinely promising on-ramp to the gospel for people who would never walk into a building (news.ag.org, Feb. 2026). But Kevin Smith, pastor of Northland Cathedral in Kansas City, raised the same caution in that same reporting: “Soon pastors will have to face the emotional connection that some lonely people, especially teenagers, make with AI applications… There is a real possibility of potential emotional addiction forming.” And D. Allen Tennison, theological counsel for the Assemblies of God, put a sharper point on it: “AI cannot pray for us or repent for sin. AI cannot mentor disciples by imitating Christ. Discipleship requires connection to flesh and blood people in church” (news.ag.org, Jan. 2026).

It would be too easy to read all of this as a simple verdict against the technology. The research itself resists that. A large panel study of Japanese adults found that companion-AI use was associated with higher well-being, life satisfaction, and sense of purpose — and the benefit was strongest precisely among people who were already lonely (ScienceDirect, Journal of Affective Disorders research literature, 2026). A pilot study published in Frontiers in Digital Health this spring found that people who chose to talk to a chatbot rather than a psychologist were, on average, more emotionally isolated to begin with — they described the bot as “a safe and non-judgmental space,” even while recognizing its “formulaic nature and limitations.” These are not people being deceived. They are people in real pain, reaching for whatever relief is within reach.

That’s the more honest starting point for a theological response: not alarm at the technology, but attention to the ache it’s responding to. Scripture does not treat human aloneness as a neutral fact to be managed. In the second chapter of Genesis, before sin, before any disorder had entered creation, God names something as “not good”: “It is not good for the man to be alone. I will make a helper suitable for him” (Genesis 2:18, NIV). Aloneness is the first thing in an otherwise “very good” creation that God identifies as deficient — not a flaw in the man, but a gap built into the structure of being human, one that only another embodied person could fill. The writer of Ecclesiastes draws out what that means practically: “Two are better than one… If either of them falls down, one can help the other up. But pity anyone who falls and has no one to help them up… A cord of three strands is not quickly broken” (Ecclesiastes 4:9–12, NIV). The comfort in view there isn’t information or conversation — it’s a hand that can physically lift you when you’re down, a body of warmth next to yours in the cold. A chatbot can simulate the sentence. It cannot be the hand.

Pentecostal ecclesiology has a specific name for what this points to: the Church as the Body of Christ. Paul’s image in 1 Corinthians is not a metaphor for emotional support generally — it describes a structure where “there should be no division in the body, but that its parts should have equal concern for each other. If one part suffers, every part suffers with it; if one part is honored, every part rejoices with it. Now you are the body of Christ, and each one of you is a part of it” (1 Corinthians 12:25–27, NIV). This is the same conviction the Assemblies of God’s Statement of Fundamental Truths makes explicit in its article on the Church and Its Mission: believers are “born of the Spirit” into a body, not merely given information about God. The Holy Spirit’s work, in AoG theology, is not primarily to transmit content — it is to baptize believers into living, interdependent, physically present community. An algorithm, however well-trained on Pentecostal doctrine, cannot be baptized into that body. It cannot suffer with you or rejoice with you, because it cannot suffer or rejoice at all.

This is also why the writer of Hebrews frames gathering with other believers not as a preference but as a discipline under active pressure: “Let us consider how we may spur one another on toward love and good deeds, not giving up meeting together, as some are in the habit of doing, but encouraging one another” (Hebrews 10:24–25, NIV). “As some are in the habit of doing” is a strange phrase to find in a two-thousand-year-old letter and a 2026 research paper at the same time, but there it is: people drifting away from embodied gathering toward something more private, more controllable, less demanding — and a writer urging them back toward each other anyway, because the alternative was never going to satisfy what they were actually hungry for.

None of this means AI ministry tools are illegitimate. An avatar that points a searching stranger toward Scripture at 2 a.m., or that helps a missionary translate a sermon into a hundred languages, can be a genuine act of digital evangelism — “technology used for the glory of God,” as Tennison put it. The danger isn’t the tool; it’s mistaking the on-ramp for the destination. A digital avatar can tell someone about the Holy Spirit. It cannot be the hands that are laid on them in prayer, the voice that prays with them through a 3 a.m. crisis, the believer who shows up at their door with a meal. Those things were never going to be automatable, because they were never meant to be efficient. They were meant to be given, person to person, the way Christ gave himself.

If you are the lonely person reaching for a chatbot at midnight because no one else seems to be awake, that reach is not shameful — it’s evidence of exactly what God built into you in the garden. The research suggests, gently but consistently, that the chatbot will not be enough. The church’s answer was never supposed to be a better algorithm. It was supposed to be a body of people who notice when you stop showing up, who carry you when you fall, and who experience your suffering as their own. If that body has gone quiet around you, or if you’ve quietly let yourself drift from it, this is the invitation to come back — not to a service, but to the strange, slow, embodied work of being known by people who can actually hold you.