UC Berkeley professor admits to using AI to edit op-ed on students’ math skills

US news | The Guardian · collected 2026-08-20 · by Sophie Sullivan
Read the original at US news | The Guardian ↗

Summary

Zvezdelina Stankova, a math professor at UC Berkeley, admitted to using AI software to help edit an op-ed on student math skills published in the San Francisco Standard. The piece criticized the university's test-blind admissions policy, suggesting that students were being admitted without adequate preparation for calculus. According to AI-detection software, 33% of the op-ed had been generated or assisted by AI, with Stankova using the software to locate documents and articles related to the initiative. The incident has sparked debate over the use of AI in academic work, with some criticizing Stankova's actions as ironic given her criticism of student underpreparedness.
Written by the local model on 2026-08-20, using this article's own text rather than the other coverage of the same event (that is the story summary below).

Signals How these are calculated →

Claims extracted
24
claim-shaped sentences
Uncertain
17%
4 of 24 hedged
Leaning
not scored
needs a local LLM pass
Publisher trust
95.2
red-flag proxy, not a credibility rating
Outlets on this story
1
Education
Narrative spread
1
articles carrying this framing
Analyzed 2026-08-20 · how these are computed

AI analysis (generated at analysis time, not now)

Story summary

Zvezdelina Stankova, a math professor at University of California, Berkeley, wrote an op-ed for the San Francisco Standard criticizing students' math skills and blaming the university's test-blind admissions policy. However, it was discovered that 33% of the article had been generated or assisted by AI using software called Pangram. The professor initially denied any involvement with AI but later admitted to using it to help edit the piece, claiming 80% of the work was done by humans and only 20% was AI-assisted.

Written for “UC Professor Admits Cheating” on 2026-08-31, grounded in this article and the 0 other(s) covering the same event.
Why this leaning score
The article's own words the score was based on. Each is quoted verbatim and was checked against the article text before being stored, so you can find it in the original.
Score +0.35 Confidence high
Leaning score +0.35 for article 1502 (high confidence, 1 verified quote) · logged 2026-08-21

Story

📰 UC Professor Admits Cheating
Education · 1 article(s) covering the same event. This is the one the site leads with.

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Compared with similar articles

This article reads unscored and hedges 17% of its claims. Each row says how that neighbour differs.
The Independent World
⚖️ Leans right 🔴 14% hedged 5 of 35 📰 publisher trust 95
“The articles describe two different incidents involving professors using AI to aid in their work, one about a math professor editing an op-ed and another about a computing academic representing himself with AI assistance at a tribunal”
The Free Press
⚖️ Leans right 🔴 11% hedged 1 of 9 📰 publisher trust 96
“Article A reports on a UC Berkeley professor's op-ed and admission of using AI to edit, while Article B discusses the regulation of AI in general and mentions Stripe's achievement of 'singularity', without any mention of the specific event described in Article A”

Publisher

US news | The Guardian · 90 article(s) · 0 correction(s) detected
SignalValueWeight
Correction rate 0.000 0.4
Uncertainty density 0.095 0.25
Assertive mismatch rate 0.000 0.35
No corrections detected for this publisher. That may mean careful reporting, or simply that nothing has been checked.

Who wrote this

Sophie Sullivan
2 article(s) here · 1 carrying a prediction
🔮 In an unsigned statement, it described a “short-term social experiment” to examine how fake polling “could enter and spread through the political information ecosystem without independent verification”.
🔮 When asked about their AI usage policy, the Standard – which published the op-ed – shared this statement by email: “While AI may assist, our expectation is that humans are behind every article we publish and take responsibility for every word.
Also by Sophie Sullivan
Nothing else under this byline is closely related to this article, so these are simply their most recent.

Topics

Berkeley Standard the Daily Californian the San Francisco Standard the University of California, Berkeley

Subjects

Stankova PERSON · 9× Berkeley GPE · 3× Pangram ORG · 2× Standard ORG · 2× Francis Luo PERSON · 1× Harvard ORG · 1× Zvezdelina Stankova PERSON · 1× the Daily Californian ORG · 1× the San Francisco Standard ORG · 1× the University of California, Berkeley ORG · 1×

Narrative

Camille Crittenden, a member of the council, said she was unaware whether faculty had any specific guidance regarding the use of AI in personal research or editorials, but said that “most AI detection tools are still quite unreliable and lack nuance” with regard to Pangram.
framing: assertive · carried by 1 article(s) · first seen 2026-08-20
🔮 When asked about their AI usage policy, the Standard – which published the op-ed – shared this statement by email: “While AI may assist, our expectation is that humans are behind every article we publish and take responsibility for every word.
2026-08-20 · US news | The Guardian
UC Berkeley professor admits to using AI to edit op-ed on students’ math skills · assertive framing

Claims (24 extracted, 4 hedged)

A math professor at the University of California, Berkeley, criticizing a “severe” math deficiency among students in an op-ed for the San Francisco Standard, admitted to using artificial intelligence to help edit the piece. asserted
professor → criticize → piece
The Standard published a 2,000-word piece by Zvezdelina Stankova last week, in which the professor said some of her math students were “five to eight years” behind and lacked a “middle school” education on fractions and basic algebra. asserted
some → publish → fractions
Stankova said the UC system’s test-blind admissions were to blame, suggesting that students who weren’t sufficiently prepared for the rigor of Berkeley’s mathematics program were admitted because a longstanding benchmark like the SAT had disappeared. asserted
benchmark → say → SAT
Over the weekend, journalists at Berkeley’s student newspaper, the Daily Californian, noticed the op-ed’s language sounded like AI. asserted
language → notice → AI
According to Berkeley sophomore Francis Luo, they ran it through AI-detection software Pangram, which claimed 33% of the op-ed had been generated or assisted by AI. uncertain
% → accord → AI
In response, Stankova admitted she had used AI software to “help edit the piece” but that the article was “the result of several hundred person-hours of intensive human work, of which about 80 hours are my own”. asserted
hours → admit → which
Stankova said by email she had used AI to locate “numerous documents and articles related to the initiative” but that “all analysis is the result of the team members”. asserted
analysis → say → members
When asked about their AI usage policy, the Standard – which published the op-ed – shared this statement by email: “While AI may assist, our expectation is that humans are behind every article we publish and take responsibility for every word. uncertain
we → ask → word
Our understanding in working with the author of this op ed was – and continues to be – that this piece reflects her and her colleagues’ extensive original analysis, research and expertise.” asserted
piece → work → analysis
Stankova’s admission has prompted a larger debate on social media on the use of AI by academics. asserted
admission → prompt → academics
Some criticized Stankova, calling its use “ironic” for a piece discussing students’ underpreparedness, or arguing an op-ed should have been entirely original and written without artificial assistance. asserted
ed → criticize → assistance
Others defended the professor, saying the use of AI was not a serious issue and distracted from the argument at hand. asserted
use → defend → hand
Stankova, in her email, said “the issue of how AI was used … is orthogonal to and a distraction from the thousands of hours our team has put into the initiative” of reinstating standardized tests in admission. asserted
team → say → admission
The debate over whether tests like the SAT or ACT should be again used in UC admissions has intensified over recent years. asserted
tests → use → years
In 2020, the UC regents voted to phase out the tests as a requirement for application; so did institutions like Harvard, Stanford, MIT and Yale. asserted
institutions → vote → Harvard
Each of those peer colleges have since reinstated the requirement. asserted
Each → reinstate → requirement
In her op-ed, Stankova shared data that she and her peers had analyzed, which suggested that the number of students with a “severe deficit” in their readiness for a calculus I class had tripled after test-blind admissions were introduced. uncertain
admissions → share → class
Prior to publishing her op-ed last week, Stankova – in addition to more than 3,000 other UC faculty members – signed a June letter supporting admissions testing. asserted
Stankova → publish → testing
Stankova, one of the letter’s authors, said AI was also used for editing there. asserted
AI → say → editing
The UC academic senate said in a late July statement that it would begin a review to determine whether standardized tests would again be used in admissions. asserted
tests → say → admissions
If reinstated, the changes would affect fall 2028 admissions at earliest. asserted
changes → reinstate → admissions
The UC system also has its own “AI council”, which helps develop AI initiatives and resources for its colleges. asserted
which → have → colleges
Camille Crittenden, a member of the council, said she was unaware whether faculty had any specific guidance regarding the use of AI in personal research or editorials, but said that “most AI detection tools are still quite unreliable and lack nuance” with regard to Pangram. asserted
tools → say → Pangram
On its website, Pangram claims it correctly identifies AI-generated documents 99.66% of the time, or gives one false positive about every 24,000 documents scanned. uncertain
it → claim → documents
💬 Give feedback
🕘 History 🎫 Support