Anthropic Surveyed 81,000 Claude Users — Hope and Alarm Coexist in the Same Person

The discourse surrounding artificial intelligence adoption has frequently been framed as a binary proposition: you are either an enthusiastic adopter embracing the future, or a cautious sceptic wary of the consequences. However, a landmark study released by Anthropic fundamentally dismantles this simplistic narrative. In what is believed to be the largest and most multilingual qualitative study ever conducted, involving 80,508 people across 159 countries and 70 languages, the findings reveal a far more complex reality.

People are not splitting into pro-AI versus anti-AI camps. Instead, hope and alarm coexist simultaneously within the same individuals. The Anthropic Interviewer study, which engaged users in conversational interviews over a week in December 2025, exposes a profound tension at the heart of our relationship with this technology. Those who perceive the greatest potential benefits from AI in a specific domain are often the very same individuals who harbour the deepest fears about its implications.

The most prominent aspiration identified in the survey, cited by 18.8% of respondents, is "Professional excellence." Users are leveraging AI to navigate complex tasks, streamline workflows, and enhance their cognitive capabilities. The narrative of AI as a powerful productivity multiplier is clearly resonating across global professional landscapes, with 32% of users reporting that the technology has dramatically sped up their work and tedious tasks.

However, this aspiration is inextricably linked to significant anxieties. The study highlights that concerns regarding "Unreliability" (26.7%) and "Jobs & economy" (22.3%) are paramount. The tension here is palpable: professionals want AI to make them better at their jobs, but they simultaneously fear that the tool they rely on might be untrustworthy—causing domino effects through inaccuracy and hallucinations—or worse, that it might eventually render their roles obsolete.

Consider the implications for project delivery professionals in the construction and AEC sectors. We are actively seeking tools to manage the overwhelming complexity of modern projects, from contract review to supply chain optimisation. Yet, as we integrate these systems, the fear of relying on flawed outputs or losing the critical thinking skills honed through years of experience remains a persistent undercurrent. The lawyer quoted in the Anthropic study encapsulates this perfectly: using AI to review contracts saves time, but it raises the fear of losing the fundamental ability to read and think independently. Conversely, the study also captured profound hope for cognitive partnership—the desire to work alongside AI as a thinking partner rather than as a replacement.

Another significant finding is the desire for "Personal transformation" (13.7%) and using AI as a collaborative entity for brainstorming and problem-solving. Users are finding value in having a tireless, knowledgeable sounding board. Interestingly, only 5.6% of respondents cited "creative expression," making it the least prioritised use case, suggesting a shift towards practical utility over mere novelty.

Yet, this desire for cognitive assistance is mirrored by a fear of cognitive atrophy. If we increasingly offload our analytical and creative processes to an algorithm, do we risk degrading our own intellectual capacities? This is a critical question for disciplines that rely heavily on nuanced judgement and strategic foresight. The convenience of an immediate, AI-generated solution must be balanced against the necessity of maintaining our own analytical rigour.

Furthermore, the study reveals a complex relationship with "Autonomy & control" (21.9%). While users want AI to empower them, they are deeply concerned about the potential loss of agency. This is particularly relevant when considering the deployment of autonomous agents in high-stakes environments. The desire for efficiency cannot supersede the requirement for human oversight and accountability.

The Anthropic study provides a necessary corrective to the often abstract projections of AI's future. By grounding the conversation in the concrete experiences of tens of thousands of users, it reveals that "AI going well" is not a straightforward trajectory. It is a nuanced negotiation of competing desires and fears.

It is also important to acknowledge the study's limitations. As industry analysts have pointed out, surveying only existing Claude users may present a skewed perspective, potentially under-representing the scepticism found in broader public surveys, such as those from Pew Research, which indicate higher levels of distrust. Furthermore, the study highlighted significant regional differences: respondents from India, Brazil, and Israel showed a largely positive outlook, while those from Germany, South Korea, and the United Kingdom were generally more sceptical.

This nuanced understanding of user sentiment is crucial for any organisation navigating the complexities of AI adoption. The assumption that providing access to powerful tools will automatically result in enthusiastic uptake is flawed. Adoption strategies must acknowledge and address the legitimate anxieties that coexist with the desire for innovation.

The complex realities of AI adoption require more than just technological implementation; they demand thoughtful integration and robust governance. If you want to hear more about how project delivery is evolving in response to these challenges, subscribe to the Project Flux newsletter.

•https://www.anthropic.com/81k-interviews

•https://www.darioamodei.com/essay/the-adolescence-of-technology

•https://indianexpress.com/article/technology/artificial-intelligence/anthropic-study-reveals-what-users-expect-from-ai-and-what-worries-them-10593924/

All content reflects our personal views and is not intended as professional advice or to represent any organisation.