OpenAI and Anthropic studied how people use ChatGPT and Claude. One big difference emerged. (www.businessinsider.com)

OpenAI and Anthropic each published large user-behavior studies this week, revealing distinct patterns in how people use their models: ChatGPT is used mainly for augmentation — drafting, editing, summarization, brainstorming and interactive help — while Claude is shifting toward automation, embedded in workflows for coding, document processing and report generation. OpenAI categorized chats as “Doing” (produce plug‑and‑play outputs), “Asking” (decision support without direct outputs) and “Expressing,” and reported that Asking dominated usage as of late June. Anthropic’s latest (its third) study found coding and math-based tasks account for over a third of Claude usage, and this edition is the first where automation overtook augmentation. The divergence matters for product strategy and enterprise adoption: ChatGPT’s strength is iterative, creative collaboration with individual users (students, marketers, knowledge workers, educators), while Claude is increasingly used to delegate whole tasks at scale inside companies. Technically, that implies different optimization priorities — conversational UX and content generation for ChatGPT versus reliable programmatic interfaces, document ingestion, and deterministic automation pipelines for Claude. Enterprises may run both: ChatGPT for ideation and communication, Claude for structured, productionized execution.