What People Actually Want From AI
The artificial intelligence industry spends considerable energy debating what models can do. Anthropic asked a different question: What do people actually want them to do? The answer, drawn from 80,508 interviews across 159 countries and 70 languages, is more personal than most industry projections suggest, and it maps closely onto the behavioral shift that consumer data is already capturing in real time.
The study, conducted through Anthropic’s AI interviewer tool and published this week, is described by the company as the largest and most multilingual qualitative study ever conducted.
Participants were asked open-ended questions about what they hoped AI could make possible and what they feared it might do. What emerged is not a picture of people seeking technological novelty. It is a picture of people trying to solve practical, familiar problems: not enough time, not enough money, too much low-value work standing between them and what they are actually trying to accomplish.
What People Said They Wanted
The largest group, 18.8% of respondents, said they wanted professional excellence, according to Anthropic’s findings. This is the ability to hand off routine cognitive work so they could focus on the parts of their jobs that required judgment and skill.
A healthcare worker in the United States described receiving 100 to 150 messages a day from doctors and nurses, with so much time consumed by documentation that little remained for patients or colleagues. Since using AI, she said, that pressure had lifted. She had more patience. More time to explain things to families.
Personal transformation came second at 13.7%, with respondents describing AI as a coach for behavior change, emotional growth and wellbeing. Life management followed at 13.5%, with people seeking help managing schedules and reducing the mental load of daily organization. A manager in Denmark said that if AI handled that burden, it would give back something irreplaceable: the ability to be fully present.
Time freedom ranked fourth at 11.1%, followed by financial independence at 9.7% and societal transformation at 9.4%.
A software engineer in Mexico said AI now lets him leave work on time to pick up his children from school. An entrepreneur in Cameroon said AI allowed him to reach professional competency in cybersecurity, design, marketing and project management simultaneously in a country where he could not afford many failures. “Finding a payment platform available in my region would have taken me a month,” he said. “AI did it in 30 seconds. It’s an equalizer.”
The concerns were concrete as well. Respondents worried about dependency, job loss, and whether AI access would be distributed fairly. Hope and alarm, Anthropic found, tended to coexist in the same person rather than divide people into opposing camps.
What Consumer Data Shows
The aspirations Anthropic captured are showing up directly in consumer behavior. AI use for personal tasks reached 54% of U.S. consumers in January, up five points from December, according to PYMNTS Intelligence. One-third of consumers now qualify as mainstream users, meaning adoption is deepening.
The most significant behavior change is where people start. More than one-third of Gen Z consumers and power users now turn to a dedicated AI platform first when tackling personal tasks, bypassing search entirely, as PYMNTS Intelligence reported.
Reliance on dedicated AI platforms increased 36% among Gen Z and 28% among power users in a single month. Among those who rely primarily on AI platforms, 42% report using search engines less, compared with 33% among those accessing AI through search summaries, according to PYMNTS Intelligence.
The average user now relies on two to three different AI tools, and 52% of power users access AI through installed apps rather than browsers, according to PYMNTS Intelligence. Claude’s user base rose 22% in January month over month while Gemini grew 8%, suggesting consumers are actively testing alternatives as the market matures, as PYMNTS Intelligence found.
Gap Between Want and Reality
Consumer data can show that people are using AI more and starting there first. It cannot easily show what they are hoping it will do when they open it. The 81,000 interviews answer that.
People are not primarily looking for a faster search engine. They want meaningful relief: less time on documentation, fewer hours on tasks that drain focus, more capacity for the work and the people that matter to them.
That distance between what people want and what current tools consistently deliver is where the next round of product competition will be decided. The respondents who were most enthusiastic were also the ones who had experienced AI working well in one situation and wanted more of it everywhere else.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
The post What People Actually Want From AI appeared first on PYMNTS.com.