Gates knows, too, that prudence and discretion might slow down or even derail that trillion dollar virtual gravy train.
"In private, people who understand how good this stuff is, and how much better it’s getting, they’re very worried." Bill Gates is scared af about unchecked, unregulated AI development. www.nytimes.com/2026/08/26/t...
— Liz Renzetti (@lizrenzetti.bsky.social) August 26, 2026 at 7:55 AM
[image or embed]
Here’s the abstract of a study just released June, 2026:
Using 30 months of panel data on 26,811 Chinese students in grades 7--12, we study how generative AI affects homework productivity and learning. The data combine monthly closed-book exams, high-school and college entrance exams, and homework scores and completion time across nine subjects. We exploit staggered AI adoption in a difference-in-differences design. AI adoption raises homework scores by 18% and reduces completion time by 30%, but lowers monthly exam scores by 20% within six months. High-stakes entrance-exam scores fall by 18 and 24%, with the full penalty emerging only after about two years. The losses are largest in social science subjects, followed by STEM and languages, and are especially large for junior students, high-achieving students, and boys. The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores. AI users who maintain similar homework completion time as non-AI users experience small learning losses.
Join the growing movement to establish a five-year moratorium on the use of generative AI in schools.
From Fairplay’s Position Statement:
As generative AI products1 proliferate in educational spaces, they are posing risk of significant harm to children. We, the undersigned, call for a five-year pause on all student-facing generative AI products in PreK-12 schools. Currently, these products threaten:
a) Student and educator privacy and autonomy;
b) Skill development, including cognition, critical thinking, analytic reasoning, decision-making, emotion regulation, and relationship-building;
c) Mental health, fairness, safety, and the right to a high-quality education;
d) Educators’ role as professionals;
e) Academic integrity;
f) The environment.
. . . .
CALL FOR FIVE-YEAR PAUSE
The rapid expansion of generative AI products into schools without oversight, community input, or evaluation of implications is not inevitable. A five-year pause on all products using generative AI that impact children in pre-K-12 schools would allow time for school communities, including students, educators, administrators, and parents, to learn about the implications and uses of generative AI products in education, to ask questions, and to provide feedback.
In particular, it would give time for schools to train staff and ensure that any generative AI products used will:
● Improve learning outcomes without cognitive offloading or impeding human relationships;
● Demonstrate absolute safety for use by students (addressing issues of addiction, persuasive design, data and privacy risks, exposure to harmful content, mental health, parasocial relationships, cyberbullying, etc);
● Not be used for non-authorized purposes such as cheating, academic
dishonesty, or plagiarism;
● Sufficiently consider and prioritize privacy, civil rights, ethics, justice, and climate impacts of generative AI products;
● Never be used in place of teachers, especially for vulnerable populations such as neurodivergent students, at-risk students, and students of low socio-economic status.
Until and unless the above can be shown (or there is evidence to support all of the above), generative AI products should not be used in pre-K-12 schools. Further, state and provincial governments and education departments, and federal, state, and provincial regulators should use this pause to develop and implement:
● An audit of existing generative AI platforms for efficacy, safety, and legality, performed by neutral, independent third parties;
● A registry of generative AI products currently in use, including the location of collected data, especially the intellectual property of students and teachers;
● A vetting process for new generative AI products prior to their introduction into pre-K-12 schools;
● A framework for culturally-responsive, relational approaches to communication that provides opportunities for technology-free, play-based learning spaces;
● Transparent, thoughtful, protective, and rigorous protections governing the use of generative AI products in schools, independently vetted by neutral third parties;
● Sufficient policies to protect student data, eliminate any advertising or marketing, prohibit addictive algorithms and gamification, and forbid products that maximize engagement or profit off student data.
Written by members of the Screen Time Action Network’s Screens in Schools Work Group: Emily Cherkin, MEd, Faith Boninger, PhD, Shaleen Title, MS, JD, Denise Champney, MS, CCC-SLP, and Kelly Clancy, PhD. No generative AI was used to produce this document.
