The biggest AI survey conducted so far uncovers what humans truly desire.

      Anthropic's extensive AI interview study goes beyond technology; it serves as the largest reflection of human desire ever created.

      In December 2025, a software engineer in Mexico wrapped up his workday early to pick up his children from school. A lawyer in India met with an AI tutor and, for the first time, read Shakespeare without hesitation. A soldier in Ukraine, restless amid shelling, opened a chat window to teach himself something new, as learning was his only means to stave off panic.

      These individuals were strangers to each other. Yet, during the same week, they each engaged in an open-ended conversation with an AI interviewer developed by Anthropic, the company behind Claude.

      They answered a seemingly simple set of questions: what do you desire from AI, has it met your expectations, and what worries you? Over the course of seven days, 80,508 individuals across 159 countries and in 70 languages participated. The outcome, released in March 2026, is described by Anthropic as the largest and most multilingual qualitative study ever conducted.

      What respondents yearned for most from AI, as categorized by Claude from their open-ended responses to “If you could wave a magic wand, what would AI do for you?” indicated that 1% of participants did not express a vision. Credit: Anthropic

      This assertion is credible. The previous records were held by the USC Shoah Foundation’s Visual History Archive, which documented approximately 52,000 to 59,000 testimonies of genocide across 40 languages, and the World Bank’s “Voices of the Poor” project, capturing experiences from over 60,000 people in 60 countries.

      Anthropic’s study surpasses both in terms of respondent count, although such comparisons must be approached cautiously due to the differing depth and significance of those earlier projects.

      The real value of the Anthropic study lies not in its size but in what the size unveils. When 81,000 individuals are posed an open question regarding their aspirations for a technology, and the most common response leans towards “give me my life back” rather than “make me more productive,” it transcends into a reflection of deeper human yearning rather than a mere product assessment.

      The calculus of desire

      An adaptive version of Claude, the Anthropic Interviewer, posed a consistent set of questions to each participant, then followed the conversation's natural flow. Claude-powered classifiers subsequently categorized every response.

      This method reconciles the traditional balance in social science between depth and breadth, achieving a speed that was unimaginable just two years prior.

      While the headline figures are informative, they do not encapsulate the complete narrative. Roughly 19% of participants sought AI for “professional excellence,” representing the largest category. Another 14% desired “personal transformation” through emotional growth or health support. Approximately 11% wanted more time for family and leisure, while nearly 10% sought financial independence.

      These neatly organized figures might seem like a product roadmap, but they are not the essence of the matter.

      What interviewers consistently discovered was that the initial response often masked a more profound truth. Individuals might begin by discussing the need to automate emails or expedite coding, yet when probed about the ultimate outcome, they expressed desires such as wanting to cook with their mothers instead of completing tasks.

      The language of productivity was borrowed from the technology itself; the underlying aspiration was more fundamentally human: relief from the mental burdens of modern life.

      This pattern is recognizable to anyone familiar with qualitative research. However, the confidence afforded by its presence across countless conversations in 70 languages, spanning from Lagos to Lyon, is unparalleled.

      Experiences beyond surveys

      Surveys indicate what people select; interviews reveal why they hesitate before making their choices. The most valuable content in Anthropic’s report lies not in the categorized data but in the quotes, which resemble memoir fragments rather than mere customer feedback.

      A butcher in Chile, who had scarcely used a computer prior to AI, spoke of embarking on entrepreneurship and discovering unexpected motivation.

      A homeless healthcare worker in the United States utilized AI to brainstorm a digital marketing business and, for the first time, glimpsed a pathway to home ownership. A physician in Israel, unable to obtain a diagnosis for her neurological condition from local doctors, leveraged AI to find two research studies that led to effective treatment. A mute worker in Ukraine created a text-to-speech bot with Claude, enabling near-real-time communication with friends.

      These are narratives focused not on productivity but on accessibility. They converge around a range of AI capabilities unrelated to speed: patience, availability, and the absence of judgment.

      A student in India noted that his professor, teaching 60 students, rarely entertained questions, whereas AI allowed him to inquire freely at 2 am. A lawyer in the same country shared her triumph over a lifelong fear of math, realizing she was not as unintelligent as she had previously thought.

      Eighty-one percent of respondents claimed that AI had already