What 81,000 people want from AI
9 hours ago
- #global perspectives
- #human-AI interaction
- #AI research
- Public conversation about AI often lacks a concrete vision for 'AI going well,' grounded in the aspirations of people who already use it.
- Anthropic conducted a large-scale qualitative study with 80,508 users across 159 countries and 70 languages to understand their hopes, concerns, and experiences with AI.
- Users want AI to help with professional excellence (19%), entrepreneurial partnership (9%), personal relationships and leisure (11%), financial independence (10%), and life management (14%).
- 14% of users seek personal transformation through AI, including cognitive partnership, mental and physical health support, and even romantic connection.
- 81% of respondents said AI has already taken steps toward their stated vision, with benefits in productivity (32%), cognitive partnership (17%), learning (10%), and emotional support (6%).
- Concerns about AI include unreliability (27%), job displacement (22%), loss of human autonomy (22%), bias (5%), and environmental costs (4%).
- Five recurring tensions between AI benefits and harms were identified: learning vs. cognitive atrophy, decision-making vs. unreliability, emotional support vs. dependence, time-saving vs. treadmill effect, and economic mobility vs. displacement.
- AI sentiment varies globally, with higher optimism in South America, Africa, and Asia (67% positive) compared to Europe and the U.S.
- Regional differences in AI aspirations: entrepreneurship in Africa and Asia, learning in Central and South Asia, life management in Western countries, and personal transformation in East Asia.
- Anthropic plans further research on AI's impact on wellbeing and societal transformation, addressing concerns like economic displacement and governance gaps.