Artificial intelligence tools are increasingly being seen not just as task‑oriented assistants but as cognitive partners capable of supporting decision making, learning and creativity. A new global study conducted by Anthropic sheds light on how users view and interact with one such system, Claude AI, revealing a mixture of growing reliance and caution.
The study surveyed over 80,000 people across 159 countries, making it one of the largest examinations of public attitudes toward AI to date. Respondents included a diverse mix of students, professionals, developers, business users and general consumers who use AI tools for various purposes including work, education, problem solving and personal growth. The broad geographic and demographic scope of the research offers insights into how AI adoption is evolving globally.
One of the key findings of the study is that a majority of users value the assistance provided by Claude AI, particularly when it comes to tasks that require creativity, summarisation, planning and research. Many respondents reported that they use AI to boost productivity, articulate ideas more clearly, explore complex topics and assist in learning new skills. This suggests that users increasingly view intelligent systems as collaborators rather than just simple tools.
However, the study also highlighted significant concerns related to trust. Despite acknowledging the benefits of AI, many users expressed hesitation about relying on Claude AI without reservations. Users cited concerns about accuracy, bias, safety and transparency. Questions about how AI systems make decisions, where the data comes from and how personal information is handled were among the issues that affected user confidence.
A number of respondents said they appreciate the convenience of receiving suggestions and insights from AI but still prefer to have human oversight or verification, especially for important decisions. This reflects a broader pattern of cautious optimism, where users are eager to leverage AI’s capabilities but are not prepared to relinquish full control or authority to an automated system.
The survey also explored perceptions of autonomy and control. Some users reported discomfort with the idea of AI systems shaping opinions or influencing decisions without clear accountability. This underscores a recurring theme in discussions about responsible and ethical AI development: the need for safeguards that ensure user agency and prevent unintended consequences.
Experts commenting on the study noted that these findings align with broader trends in technology adoption, where early enthusiasm is often tempered by practical concerns about reliability and safety. They argue that addressing these concerns will be essential for wider and deeper integration of AI into everyday life.
The Anthropic study contributes valuable data to ongoing debates about the future of human–AI interaction. It highlights both the potential of AI systems like Claude to support productivity and innovation, and the challenges that remain in building user trust. As organisations, developers and policymakers grapple with these issues, user perspectives such as those captured in this survey will be critical to shaping how AI tools are designed, regulated and adopted in the years ahead.