Berkeley: GenAI Increases Workload

🔑 Enhanced Key Takeaways

• •The eight-month ethnographic study at a 200-employee US tech company involved in-person observations two days a week, monitoring internal communications, and 40 interviews across engineering, product, design, research, and operations.[1][2][3]
• •Workers experienced task expansion where product managers wrote code, designers handled engineering tasks, and employees tackled previously outsourced or avoided responsibilities due to AI filling knowledge gaps.[1][2][4]
• •AI blurred work-life boundaries as employees sent prompts during lunch, before meetings, or evenings, dissolving natural pauses and leading to continuous work involvement.[1][3][4]
• •Increased multitasking included running multiple AI agents simultaneously, writing code while AI generated alternatives, and managing multiple threads, creating hidden cognitive load and self-reinforcing cycles of higher speed expectations.[2][3][4]

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline