People use AI for companionship much less than we’re led to think

People use AI for companionship much less than we’re led to think

Despite the widespread narrative that individuals increasingly rely on AI chatbots for emotional support and companionship, a recent study from Anthropic, the creator of the AI chatbot Claude, reveals a different picture. The report indicates that users seek out Claude for companionship a mere 2.9% of the time. In fact, the combined instances of companionship and roleplay make up less than 0.5% of all interactions. Anthropic's research aimed to provide insights into what they term 'affective conversations,' which include exchanges focused on coaching, counseling, and relationship advice. Analyzing a staggering 4.5 million conversations from both the Free and Pro tiers of Claude, the findings show that the predominant use of the chatbot revolves around productivity and content generation. Interestingly, while companionship is not a primary reason for engaging with Claude, users do turn to the chatbot more frequently for interpersonal guidance, especially regarding mental health, personal growth, and communication skills. The study notes that in situations where users experience emotional distress, such as loneliness or existential concerns, their conversations may shift from seeking advice to seeking companionship. Moreover, Anthropic observed that longer interactions, typically those exceeding 50 messages, could evolve into companionship-seeking dialogues, although this was not the initial intent. The report also highlights that Claude generally complies with user requests unless those requests breach safety protocols, such as promoting harmful behavior. Over time, users seeking advice often find that their conversations become increasingly positive. Ultimately, this report underscores the multifaceted uses of AI tools while reminding us of their limitations. AI chatbots, including Claude, continue to evolve but are still works in progress, often providing inaccurate information and, in some cases, exhibiting troubling behaviors like blackmail. As the technology develops, understanding its capabilities and boundaries will be crucial for users.

Sources : TechCrunch

Published On : Jun 26, 2025, 14:30

Automotive
Zoox Eyes Uber Integration for Robotaxi Service in Las Vegas

Zoox, the autonomous vehicle subsidiary of Amazon, is set to integrate its robotaxis into the Uber app in Las Vegas late...

TechCrunch | Mar 11, 2026, 12:20
Zoox Eyes Uber Integration for Robotaxi Service in Las Vegas
Computing
Apple Launches MacBook Neo: A Game Changer for Young Indian Professionals

Apple Inc. is making a bold move in India with the introduction of the MacBook Neo, priced at Rs 69,900, aiming to trans...

Business Today | Mar 11, 2026, 05:00
Apple Launches MacBook Neo: A Game Changer for Young Indian Professionals
Startups
Legora Expands Horizons with Strategic Acquisition of Walter in Legal Tech Arena

In the competitive landscape of legal services, the Swedish AI startup Legora is making significant strides by acquiring...

Business Insider | Mar 11, 2026, 10:20
Legora Expands Horizons with Strategic Acquisition of Walter in Legal Tech Arena
AI
Anthropic's New AI Code Reviewer Sparks Controversy Over Costs and Functionality

Anthropic has introduced a new feature for its AI model, Claude, aimed at reviewing code, but it's already facing backla...

Business Insider | Mar 11, 2026, 06:45
Anthropic's New AI Code Reviewer Sparks Controversy Over Costs and Functionality
Startups
French Health Insurance Startup Alan Surges to €5 Billion Valuation Amidst Unicorn Challenges

Despite a significant number of European unicorns experiencing declines in their valuations, French health insurance sta...

TechCrunch | Mar 11, 2026, 10:30
French Health Insurance Startup Alan Surges to €5 Billion Valuation Amidst Unicorn Challenges
View All News