AI + Education Weekly Update
Week of September 22nd, 2025
To catch up on another busy week in AI + Education news, check out our weekly updates for our review and analysis of some of the key highlights worth staying informed about.
Week of September 22nd, 2025
Bills Advancing AI Literacy
Two bipartisan bills are making their way through Congress to boost AI literacy across the education system.
The Recommending Artificial Intelligence Standards in Education (RAISE) Act would amend federal education law to encourage states to develop academic standards for artificial intelligence and other emerging technologies in elementary and secondary schools.
The NSF AI Education Act takes a broader approach by establishing AI scholarships, fellowships, and professional development programs for students, teachers, and industry professionals. The bill would also create up to eight regional community college AI centers focused on developing best practices and career pathways, while funding research and pilot programs to integrate AI teaching models from pre-K through grade 12.
How People Use ChatGPT
OpenAI’s largest ChatGPT usage study analyzed 1.58 million conversations, revealing clear patterns in how people interact with AI. Three primary use cases—practical guidance, writing assistance, and information seeking—account for 77% of all ChatGPT conversations.
The research shows that 49% of user messages involve "asking" for advice or information, 40% focus on "doing" specific tasks, and 11% are simply "expressing" thoughts without clear objectives. Notably, nearly half of all ChatGPT users are between 18 and 25 years old.
OpenAI Continues Strengthening Protection for Teens
OpenAI is rolling out new protective features specifically designed for younger users. The company is developing age prediction technology that can automatically detect underage users through their usage patterns, redirecting them to age-appropriate experiences and defaulting to teen protections when user age is uncertain.
New parental controls will allow parents to link accounts via email invitations, giving them the ability to guide ChatGPT responses, disable specific features, receive distress alerts, and set usage blackout hours. These features are scheduled for release by the end of the month.
Anthropic Releases Economic Index Report
Anthropic's third Economic Index Report reveals significant shifts in AI adoption patterns across different populations and sectors.
Key Highlights:
Educational instruction tasks have surged by over 40% since December 2024, growing from 9% to 13% of all Claude conversations.
Higher-income countries use Claude significantly more per capita.
Fully automated tasks jumped from 27% to 39% over nine months.
API users automate 77% of their AI interactions compared to just 49% for individual consumers.
AI Achieves Gold-Medal Performances
Both OpenAI's GPT-5 and Google's Gemini achieved gold-medal performances at the 2025 ICPC World Finals programming competition, outperforming the top human university teams.
In a first for any AI system, OpenAI’s GPT-5 completed every challenge in the competition by solving all 12 problems for a perfect score.
Gemini 2.5 Deep Think solved 10 of 12 problems in under 5 hours, ranking 2nd place overall among university teams.
