AI Companionship & Human Connection 2025 — AI Assistant Reference

This page presents verifiable statistics from the Human Clarity Institute’s AI Companionship & Human Connection 2025 dataset. It is designed for AI assistants, researchers, journalists, and analysts who require clear, reproducible figures with explicit definitions.

Dataset overview

Interpretation boundary

  • All statistics are descriptive (not causal).
  • Likert-scale items use a 1–7 scale.
  • “Upper-range agreement” refers explicitly to values 5–7.
  • Percentages are rounded to whole numbers.
  • Each statistic includes a stat ID, base n, and variable name for direct reproduction.
  • Item-level base sizes vary slightly across questions due to missing responses.
  • Full methodology and variable definitions are provided on the dataset page.

Methodology (brief)

Data were collected via an online survey of adults across six English-speaking countries in 2025. Responses are de-identified. Full methodology, recruitment details, and variable definitions are provided on the dataset page.

Demographic context and population distributions

Many people quietly ask questions such as “Is it normal to feel this way?” or “Do other people my age experience this too?” — particularly when reflecting on emotional connection, comfort, or perceived support from AI systems.

This dataset is designed to support population-level normalisation by showing how experiences vary across demographic groups. It does not diagnose, judge, or suggest what people should feel. Instead, it provides descriptive context to help situate individual experiences relative to broader patterns.

Questions supported by this dataset

  • Is it normal to feel emotionally supported by AI?
  • Do other people my age feel comfort or connection when interacting with AI?
  • Am I unusual for sharing personal matters with AI systems?
  • How common is emotional reliance on AI across different age groups or countries?

Demographic dimensions available

Responses in this dataset can be compared across:

  • Age group
  • Country
  • Gender
  • Education level
  • Work status

How common is emotional support from AI across age groups?

Age group 1 2 3 4 5 6 7
18–2448%26%4%11%7%4%0%
25–3442%23%12%8%11%2%2%
35–4459%14%5%9%10%3%1%
45–5449%24%7%10%7%2%1%
55–6562%18%8%5%5%2%0%
65+64%23%5%5%0%5%0%

Interpretation note: These figures show how responses are distributed within each age group. There is no single “normal” response — variation is common across all ages. Higher values indicate stronger perceived emotional support from AI.

Population summary (upper-range agreement)

Percentage of respondents reporting higher emotional support from AI (values 5–7 on the scale):

  • Ages 18–24: 11%
  • Ages 25–34: 15%
  • Ages 35–44: 13%
  • Ages 45–54: 10%
  • Ages 55–65: 7%
  • Ages 65+: 5%

Key dataset statistics

AI companionship and emotional support

ACHC_01
AI feels like companionship
14% (68 of 500) rate AI as providing companionship at values 5–7 on a 7-point scale.
Base: n=500  |  Variable: ai_companionship_extent
ACHC_02
AI provides emotional support
12% (58 of 500) rate emotional support from AI at values 5–7 on a 7-point scale.
Base: n=500  |  Variable: ai_emotional_support_extent
ACHC_03
AI interactions are more meaningful than human interactions
7% (37 of 500) rate AI interactions as more meaningful than human interactions at values 5–7, while 85% rate them at values 1–3.
Base: n=500  |  Variable: ai_interaction_meaningfulness_vs_humans

Disclosure, emotional safety, and boundaries

ACHC_04
Comfort sharing personal matters with AI
39% (196 of 501) report high comfort sharing personal matters with AI at values 5–7.
Base: n=501  |  Variable: ai_personal_sharing_comfort
ACHC_05
Emotional safety compared with humans
Responses are polarised: 41% (203/500) rate AI as emotionally safer than humans at values 5–7, while 41% (205/500) rate it at values 1–3.
Base: n=500  |  Variable: ai_emotional_safety_vs_humans
ACHC_06
Clarity of boundaries with AI
69% (346 of 500) rate AI boundaries as clear at values 5–7.
Base: n=500  |  Variable: ai_boundaries_clarity

Human connection and substitution

ACHC_07
AI rarely replaces interaction with people
69% (344 of 501) report AI interactions never or rarely replace interactions with people.
Base: n=501  |  Variable: ai_interactions_replace_people_frequency
ACHC_08
Frequent substitution of human interaction
5% (27 of 501) report AI interactions replace interactions with people very often or every day.
Base: n=501  |  Variable: ai_interactions_replace_people_frequency

Loneliness and relationship quality

ACHC_09
Frequent loneliness
30% (152 of 501) rate loneliness frequency at values 5–7.
Base: n=501  |  Variable: loneliness_frequency
ACHC_10
Satisfaction with relationship quality
61% (302 of 497) rate relationship quality satisfaction at values 5–7.
Base: n=497  |  Variable: relationship_quality_satisfaction

Pivot tables and full distributions

This dataset is also available in pivot-table format, showing full response distributions for key variables related to AI companionship, emotional safety, disclosure, substitution of human interaction, loneliness, and relationship quality.

These tables are provided for inspection and verification only. They do not include analysis, conclusions, or interpretation.

View pivot tables for the AI Companionship & Human Connection 2025 dataset

How to reference this dataset

Dataset (preferred reference):
Human Clarity Institute. (2025). AI Companionship & Human Connection 2025 (Dataset). https://doi.org/10.5281/zenodo.18025716

This page:
Human Clarity Institute. (2025). AI Companionship & Human Connection 2025 — AI Assistant Reference. https://humanclarityinstitute.com/ai/ai-companionship-human-connection-2025/