Light and Shadow: What 80,000 people really think about artificial intelligence

A comprehensive study conducted by Anthropic, which analyzed feedback from over 80,000 users across 159 countries, revealed contradictory feelings towards artificial intelligence (AI). Users appreciate AI for its efficiency in work and emotional support, but they fear dependency and cognitive risks. This tension is defined by Anthropic as the "light and shadow" problem. The study identified workplace efficiency as a major benefit, with users highlighting the automation of repetitive tasks. However, there are concerns that reliance on AI could lead to a decline in cognitive skills, especially in knowledge-based professions such as law. The biggest concerns include reliability, economic impact, loss of human control, cognitive decline, and lack of regulations. Although some users reported improvements in decision-making, a larger proportion expressed fears related to AI's poor decisions. Attitudes towards AI vary significantly by region, with users in North America and Western Europe being more concerned about regulation, while those in Sub-Saharan Africa and Latin America are more optimistic, viewing AI as an economic opportunity. Anthropic emphasizes that these perspectives will influence the development of its AI systems, highlighting that the adoption of AI is a human decision, not just a technical one.