Overview

• Student use has normalized across high school and university, with Pew reporting U.S. adolescent ChatGPT use rose from 13% in 2023 to 26% in 2024 as educators adopt the tools more cautiously and policies lag.
• In response to integrity concerns, institutions are shifting toward oral defenses, in-class writing and transparent AI-use rules, while Turnitin says about 1 in 10 submissions contained at least 20% AI-generated text, a threshold that can reflect permitted support.
• Evidence on learning remains mixed: a UPenn study found unrestricted GPT-4 access hurt exam scores, but tutor-style prompts removed the gap, and a World Bank program combining human and GPT tutoring produced significant gains over a short period.
• Global labor research by the World Bank and ILO finds 15–34 year-olds in Latin America most exposed to AI-driven task change, reinforcing calls to prioritize human strengths such as curiosity, creativity, analytical thinking and critical judgment.
• Tech’s AI buildout is escalating, with Microsoft, Alphabet, Amazon and Meta planning roughly $300 billion in 2025 infrastructure spending as data centers powered by Nvidia chips drive heavy energy demand and raise questions about profitability and social costs, while the 70th anniversary of the Dartmouth coinage underscores how the term “AI” has long fueled public misconceptions.