Anthropic Economic Index: Insights from Claude 3.7 Sonnet (www.anthropic.com)

Anthropic has released its second research report from the Economic Index, analyzing how the launch of Claude 3.7 Sonnet—a new AI model with enhanced coding abilities and a novel “extended thinking” mode—is impacting labor market usage patterns. Based on data from 1 million anonymized Claude.ai conversations, the report highlights modest but notable increases in AI use within coding, education, science, and healthcare sectors. Extended thinking mode, designed for tackling complex problems, is predominantly engaged for technical and creative tasks by computer scientists, software developers, multimedia artists, and game designers, underscoring the model’s growing role in facilitating advanced problem-solving workflows. The study also delves into how users balance AI as an augmentative tool versus an automative one. Augmentation—which involves iterative human-AI collaboration—accounts for 57% of usage overall, with particularly high augmentation in community and social service tasks, including education. Conversely, automation, where the AI completes tasks with minimal human input, is more common in specific roles such as translators. Moreover, Anthropic introduces a first-of-its-kind bottom-up taxonomy derived from real usage patterns, identifying 630 granular task categories beyond conventional labor taxonomies like O*NET. This dataset sheds light on nuanced and emerging AI applications, from plumbing help to battery technology guidance, providing a rich resource for researchers tracking AI’s evolving economic impact as capabilities and adoption continue to grow.