Stanford HAI's ninth annual AI Index report, released April 13, finds frontier model performance between the US and China within 2.7 percentage points — a gap that has essentially closed since 2023 — while global AI investment reached a record $581 billion in 2025. Generative AI achieved 53% population adoption within three years, faster than the PC or the internet, but documented AI safety incidents rose to 362 from 233 in 2024. The report flags a "jagged frontier": leading models can win International Mathematical Olympiad gold medals but read analog clocks correctly only 51% of the time, and a 50-point trust gap separates expert opinion (73% positive on AI's workplace impact) from the general public (23%).