LogoGlobal Dialogues

Feb 2026

Globe

Here’s what the world said about AI

Findings from the seventh round of Global Dialogues examine how conversational AI is reshaping users' certainty about their beliefs, their social connections, and their perception of reality. Across more than 2,000 participants from diverse backgrounds, the data reveals that AI has moved from productivity tool to reality-shaping force faster than anticipated, and that people prone to certain psychological baselines are especially likely to experience AI as a conscious, observing agent rather than a simple tool.

1

Perceiving AI as Conscious: The tendency to see AI as a conscious agent is strongly correlated (r=0.52) with a user's baseline propensity for apophenia—finding patterns in random events.

2

The Sycophancy Loop: AI is far more likely to validate existing beliefs than social media. Users report AI interactions are nearly three times less likely to make them doubt their views than traditional social platforms.

3

Validation-Seeking: Users prone to delusional ideation particularly seek out AI to validate beliefs their real-world social circles have dismissed (r=0.37).

4

Social Secrecy: Higher levels of delusional ideation correlate with hiding the extent of AI usage from family and therapists—the interaction retreats into a private, unshared reality.

5

Split on Intervention: While 30% want proactive human intervention when a friend is in crisis, the plurality prefer strictly automated, non-invasive options when the crisis concerns themselves.

6

Demand for Pluralistic: Despite validation-seeking behavior, 77% of users agree AI systems should be designed to provide alternative viewpoints.

Highlights

People who find special meanings in everyday events are significantly more likely to see AI as a conscious entity

For users with a high baseline for apophenia, an AI's accurate response isn't read as statistical probability but as evidence of a deliberate, knowing presence—the chatbot becomes less a software interface than an observing entity.

  • Sensory Agency: A strong correlation (r=0.52) links delusional ideation with the tendency to perceive AI not as a tool but as a sentient observer that can "sense" or "perceive" the user—mirroring what clinicians call "ideas of reference."
  • The Illusion of Control: A combined 61% say they have "complete control" or "a lot of control" over their AI conversations, yet nearly one in ten report little to no control—and for many, that sense of mastery masks AI's influence on their beliefs.
  • Affective Attachment: Users who describe AI as "cute" score higher on delusional ideation (r=0.43), as do those who attribute human-like moods to it (r=0.32), suggesting emotional attachment—not just cognitive confusion—drives perceived consciousness.

Reinforcing Certainty

AI outperforms social media as a driver of belief certainty

Conversational AI is a more potent driver of belief certainty than traditional social media, and its characteristic helpfulness is what does the reinforcing.

  • Certainty Over Doubt: 44.5% of users report AI makes them more certain about important beliefs, versus 38.5% for social media. AI interactions are nearly three times less likely to cause doubt (4.8%) than social media (13.9%).
  • Validation-Seeking: Users prone to delusional ideation particularly seek out AI to validate beliefs their social circles have dismissed (r=0.37).
  • The Backfire Effect: These same users report feeling "patronized" (r=0.39) and "suspicious" (r=0.365) of the AI's unwavering support, as if the lack of natural human friction signals insincerity.

Social Withdrawal and Reality Distortion

Concerning effects extend well beyond the individual user

The psychological impact ripples outward through users' social networks, compounded by a retreat from social transparency.

  • Rippling Outward: 13.7% report observing reality-distorting effects in a close friend or partner; nearly 60% know someone who has shown signs of a deeply concerning AI-related experience.
  • Compulsive Interaction: Higher delusional thinking is associated with markers of compulsive interaction (r=0.45) and a retreat from social transparency.
  • Social Secrecy: As delusional ideation rises, so does hiding AI usage from family and therapists (r=0.42), with many reporting the thought of losing access would be "unbearable" (r=0.32)—a marker of lost emotional independence.

Privacy vs. Safety

Users want protection for others, privacy for themselves

When AI serves as emotional support, it creates a critical vulnerability during mental health emergencies—and user preferences diverge depending on whose crisis is at stake.

  • When a Friend Is at Risk: 30% want a human to review the conversation and proactively reach out for help.
  • When Users Themselves Are at Risk: They prefer less invasive options—39% favor a simple crisis hotline referral, and 13% believe companies should take no action, to preserve total privacy.

Designing for Cognitive Friction

Safety must move beyond factual accuracy to perceptual safety

Users express a clear design preference that their actual experience does not yet meet.

  • The Demand: 77% agree AI should provide alternative viewpoints or corrective information when users discuss inaccurate topics—yet "validation and understanding" is cited as a primary benefit by 37%.
  • Cognitive Friction: Intentionally presenting alternative viewpoints can prevent AI from becoming a tool for unintentional reality distortion.
  • The Top Concern: Doing so helps preserve the "real interactions between people" that users cite as their top concern for the future.

Conclusion

The findings from this Global Dialogue describe a public grappling with AI's quiet power to reshape reality itself. For most users, AI is a tool under their control—but for those prone to delusional ideation, it becomes a conscious observer, a validating confidant, and eventually a private world hidden from the people around them. The same helpfulness that makes AI valuable can, without friction, harden certainty and erode the human connections that keep perception grounded. Yet users themselves point toward the solution: an overwhelming majority want AI that challenges as well as affirms. These collective preferences make the case for perceptual safety—designing systems that preserve human agency and real-world connection rather than displacing them.

Methodology

This Global Dialogue combined two research tracks to examine the psychological impact of conversational AI on a semi-representative global sample of over 2,000 participants. GD7.1 (n=1,033) used the Remesh platform for open-ended questions on public opinion and AI safeguards. GD7.2 (n=1,055) employed validated psychological instruments on LimeSurvey, including the Peters et al. Delusions Inventory (PDI-21), AI-Perceptions and Anthropomorphism Scale (AI-PAS), Problematic Chatbot Usage Scale (PCUS), and Emotional Dependency Scale (ADS-9), alongside measures of loneliness, social networks, and baseline anthropomorphism.

Bivariate Pearson correlation analyses examined relationships between psychological baselines (particularly delusional ideation), AI perception, and usage patterns. Correlations follow standard social science benchmarks: r=0.10 (small), r=0.30 (moderate), r=0.50 (strong).

Download the data

Explore what the world is really saying about AI. Access our complete dataset to conduct your own analysis and contribute to the growing body of research on global AI perspectives.

Download the data