When you ask San Francisco students whether they’re allowed to use AI for schoolwork, they each have a different answer.
“It varies even from teacher to teacher,” said 13-year-old Hayzen Chan, who goes to Presidio Middle School. “I have a teacher who lets us use AI grammar checkers, and I have some teachers who just say no AI ever.”
Chan wishes there was a clear policy. “Once a teacher says that you can use some AI, there's some kids who are bound to just take that as a free-for-all.”
He says he doesn’t use AI for schoolwork, but he thinks most of his classmates do — not to write for them, just to double check their work.
In San Francisco, the rules are also different from school to school.
Antoine LaGarde, who teaches at San Francisco International High School, said students there have to submit first drafts on paper so teachers can verify it’s original work.
Lowell High School’s 2026-27 artificial intelligence policy is available online. Ironically, it was revised with the help of artificial intelligence. Students there are allowed to use AI to complete work, but only if the teacher explicitly permits it and the student cites the use. Sometimes that means linking to an entire transcript of their conversation with a chatbot.
KALW requested official AI policies from all the district’s traditional high schools, but the only one that responded was Ruth Asawa School for the Arts. Their students “must complete their own academic assignments and examinations” and the policy prohibits “unauthorized technology.” But the school did not clarify if AI platforms qualify as “unauthorized technology.”
About 70% of teens report using AI to help with their schoolwork, according to Common Sense Media.
A district representative did not answer questions about why the district has gone this long without a wider policy, or how it’s advising schools in the meantime.
The district says generative AI programs are blocked on school-issued computers. But Chan at Presidio Middle said that’s not the case, and even double checked his access during a phone interview.
“You can still use Claude,” said Chan. “They blocked ChatGPT, but besides that I don’t know any other AI sites that are blocked.”
This debate extends beyond schoolwork. Some parents only want their kids learning from humans, and worry about AI tools like Amira already integrated in some classrooms. And teachers are worried about job security. AI was on the teachers union’s radar last year, when they negotiated a new contract for their nearly 6,000 members. The agreement they won says that AI cannot be used to replace or reduce their work, or “evaluate, monitor, or discipline” them. But they know that as their students grow up alongside this technology understanding how to use it will matter.
The district is in the process of writing its AI policy by February. The school board will vote on it on March 9, 2027.
