ASUC legislation increasingly flagged for AI use

Ahana Sur | File

Note: This data piece is part of a larger collection of stories about AI use in Berkeley. Read the articles here.

Detected AI use across a wide range of official communications and documents has steadily increased over the past several years. Using a model from Pangram, an AI detection lab, The Daily Californian scanned more than 1 million words from more than 1,ooo documents.

The data displayed a sharp increase in detected AI writing, particularly in the past year for select test groups.

Pangram claims that, on average, for every 10,000 pieces of human-written text, one will be incorrectly flagged as AI-generated. External studies, including one at the University of Chicago, found that Pangram achieved near-zero error rates. However, the software can still produce false results.

Pangram breaks a document into sections and sorts it into three categories: human, mixed and AI-generated. Text falls under the AI-generated category if Pangram deems at least 80% of its characters to have been generated by AI, while text believed to be at least 90% human-written is labeled as “Human.” Anything in between is given a “Mixed” rating.

AI writing was detected by Pangram in ASUC legislation dating back as far as the 2022-23 session, including in 47 ASUC Senate bills that were reported as being fully generated by AI. Of the legislation detected by Pangram as being fully AI-generated, 34 came from just the last 2025-26 ASUC Senate session.

One bill, Senate Resolution 2025/2026-147, endorsed and directed the transmission of a letter from Academic Affairs Vice President Dylan Lucks to the UC Board of Regents, urging them to conduct a rigorous review of the question of reinstating standardized testing requirements for first-year applicants to UC campuses. The bill itself and the associated letter were flagged as being 100% and more than 90% AI-generated, respectively, by Pangram. Lucks referred the Daily Cal to a statement by ASUC President Margaret Solomon on behalf of the ASUC executives. In an email, Solomon stated, “AI can only be a tool, not a crutch; we always write our own resolutions.”

Several ASUC senators reached by the Daily Cal for comment maintained that AI was only ever used as an editing tool.

“If our office, at any point, utilized AI, it was only used in an editing capacity or to ensure that the general resolution structure remained consistent,” said 2025-2026 ASUC senator Nicole Nuñez-Rivera in an email.

A scan of 100 official communications uncovered several emails sent by campus to students that appear to be at least partially written by AI; flagged emails, however, constituted a minority of all emails scanned. These included an announcement of upcoming changes to the campus add/drop deadline, an annual reminder on emergency response and evacuation procedures and a letter of retirement from Stephen C. Sutton, former vice chancellor for student affairs. Each was flagged by Pangram as containing more than 50% content written with AI.

Additional scans run by the Daily Cal also flagged a video speech, newsletters from campus administrators and campus departmental newsletters. Among the flagged communications were three obituaries: one from the Theater, Dance, and Performance Studies department, one from the Electrical Engineering and Computer Sciences department and one from Cal Performances.

In an Instagram post last month, Berkeley Mayor Adena Ishii shared a video, the transcription for which was flagged as fully AI-generated by Pangram. When questioned by the Daily Cal about the accuracy of the claim that the entire script was AI-generated, Ishii’s director of communications, Fran Weems, provided a statement: “We pride ourselves on authentic and transparent communication. Like any team, we use modern tools that are at our disposal.”

Copyright © 2026 The Daily Californian, The Independent Berkeley Student Publishing Co., Inc.