A new study from Anthropic suggests that people around the world are increasingly relying on AI tools like Claude to help with work, learning, and everyday life. In fact, for many, AI chatbots are becoming like a ‘cognitive partner’, not just helping them follow prompts but also actively participating in a user’s intellectual and creative processes. However, even with this increasing use, many users are still not ready to fully trust the technology.
The latest findings from Anthropic come from what the company calls its largest interview-based research project so far. In this study the company surveyed more than 80,000 users across 159 countries who shared their experiences and expectations from AI. According to the company, the conversations were conducted using a version of Claude itself, making it one of the largest and most multilingual qualitative studies ever done on how people actually feel about AI.
The research was carried out over a week in December, when all Claude.ai account holders were invited to take part in a conversational interview about what they want from AI, what worries them, and how the technology is already affecting their lives.
People want Claude to improve their lives
One of the clearest takeaways from the study is that people want AI to improve their lives, not just make them more productive. Many respondents said they use AI to handle routine work, organise their schedules, or help them learn new skills, but the deeper goal is often to gain more time, reduce stress, or focus on things that matter outside work. Anthropic noted that many answers pointed to a shared desire for AI to support personal wellbeing, creativity, and growth rather than simply speeding up tasks.
But there are trust issues
While people are increasingly depending on AI, there is also hesitation. The study shows that enthusiasm for AI comes with concerns. Users raised a wide range of worries, including unreliable answers, job losses, loss of control, privacy risks, and the possibility of becoming too dependent on AI systems. In many cases, people said they appreciated the help AI provides but still felt unsure about trusting it completely, especially for important decisions. Anthropic’s analysis found that worries about accuracy and reliability were among the most common concerns, followed by fears about the economic impact of AI and the risk of losing human autonomy.
Meanwhile, some of the responses also highlighted how emotionally complex the relationship with AI can become. One participant said they had started sharing personal problems with AI instead of talking to friends, only to later regret it, saying the choice damaged a real-life relationship.
Others described positive experiences that show why AI has quickly become part of daily life. A lawyer from India said AI helped them overcome a long-standing fear of mathematics and literature by explaining concepts in simple language, while another user in the United States said AI helped connect medical information in a way that led to a correct diagnosis after years of confusion.
Overall, the study found that people are not simply optimistic or pessimistic about AI. Instead, most users feel both at the same time, hopeful about what the technology can do, but worried about what could go wrong.
