This academic paper examines the potential macroeconomic impacts of artificial intelligence, particularly generative AI, focusing on productivity, wages, and inequality. It uses a task-based model to estimate that while AI will likely improve productivity through automation and task complementarities, the gains in the next decade will be modest, possibly less than previously predicted. The analysis further suggests that AI's effect on wage inequality will be limited, potentially widening the gap between capital and labor income, and that newly created AI tasks could have negative social value, impacting overall welfare. The author argues that realizing more substantial and positive economic benefits from AI necessitates a focus on creating new, valuable tasks for human workers, an area currently receiving less attention.