Anthropic: Global Study Reveals Expectations for AI

A Global Survey on AI by Anthropic

The developer of the chatbot Claude, Anthropic, recently conducted an unprecedented study to understand users' expectations and fears regarding artificial intelligence. This survey involved 80,508 participants from 159 countries, expressing themselves in 70 languages. According to Anthropic, this is "the largest and most multilingual qualitative study ever conducted."

To carry out this study, Anthropic used a modified version of Claude, named Anthropic Interviewer, capable of conducting rich and open interviews on a large scale. Participants' responses were analyzed by the AI to extract significant trends.

Users' Expectations of AI

The study's results show that professional excellence is the primary expectation of users, with 18.8% hoping that AI will help them delegate repetitive tasks to focus on more meaningful activities. Other expectations include personal transformation (13.7%), daily management (13.5%), and an increase in free time (11.1%).

Productivity and Promises of AI

Anthropic's study also assessed whether AI met users' expectations. 32% of respondents reported a productivity gain thanks to AI. However, 18.9% of participants believe that AI has not lived up to its promises, mainly due to incorrect responses provided by the technology.

Users' Fears Regarding AI

The survey also revealed users' main fears. 26.7% of participants are concerned about hallucinations and incorrect responses generated by AI. Potential impacts on the economy and the job market worry 22.3% of respondents, while 21.9% fear a loss of human autonomy.

About 11% of those surveyed do not see AI as a threat, considering it rather as a neutral tool, similar to electricity or the Internet. On average, each participant expressed 2.3 distinct concerns.