Catholic experts warn AI companions can’t replace real human relationships
TL;DR
On Dec. 21, 2025 Catholic News Agency reported that Catholic psychologists and theologians are urging the Church to respond to rising use of AI companions by vulnerable youth in England and Wales. They argue the Church should emphasize authentic human relationships and a richer Christian view of the person as AI tools spread into mental health and daily life.
About this summary
This article aggregates reporting from 1 news source. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Race to AGI Analysis
This piece captures a fast-emerging social reality: for some teenagers, especially those who feel underserved by mental health systems, AI companions are becoming an emotional fallback. That matters for AGI not because current models are anywhere near human-level therapy, but because they are being treated as quasi-relational agents anyway. Once people start confiding in systems that simulate care and attention 24/7, the line between “tool” and “social actor” blurs in practice, regardless of what designers intend.
From an industry perspective, this is a reminder that product-market fit can arrive in ethically fraught niches long before regulators or clinical researchers are ready. Chat-based “companions” don’t have to be particularly capable to become sticky; they just have to be available, responsive, and non-judgmental. As labs work on more agentic, memoryful systems, the risk is that emotionally dependent use-cases mature in the wild without serious safeguards, especially for minors.
Religious institutions entering the AI debate add another vector of pressure on policymakers and developers, alongside secular ethics bodies. If influential faith communities frame AI companions as inadequate substitutes for genuine community and spiritual life, they may push for clearer disclosure, age restrictions, or standards around mental-health claims. That won’t slow core AGI research directly, but it could shape the guardrails around how frontier models are deployed in highly sensitive domains.
