According to Berkeleytime, 35.3% of CS 10 students and 10.6% of CS 61A students received F’s in spring 2026. In spring 2025 and spring 2024, the percentage of F’s did not exceed 10% for either class. The electrical engineering and computer sciences department’s grading guidelines state that 7% of students in lower division courses, including CS 10 and CS 61A, should receive D’s and F’s.
UC Berkeley teaching professor Dan Garcia taught both CS 10, “The Beauty and Joy of Computing,” and CS 61A, “The Structure and Interpretation of Computer Programs,” in spring 2026. Garcia believes the “primary driver” of these abnormally high failing rates is due to a “vast increase in academic dishonesty” due to students’ usage of large language models, such as Claude, ChatGPT and Google Gemini.
It is still not entirely clear to me what happened with CS 10 although the large number of students failed for cheating seems to have been a biggish scandal (I had already read about it on reddit), but Ranade’s comments about EECS 127 (where she failed 16.8% of students) were news to me,
Ranade said students are expected to enter the course having taken classes on linear algebra, vector calculus and mathematical proofs. However, she found out in office hours that many students struggled with linear algebra, and was even more shocked when one student told her the linear algebra class they took at UC Berkeley had an “open-internet, open-AI policy” for homework and exams.
I am not sure which linear algebra course she was referring to, because my understanding is that Math 54 is a pretty tight ship!
It doesn’t seem tenable for some classes at a given school to allow AI and then have some that don’t (i.e. prof choice).…and this seems like a common policy right now.
I don’t have any answers, but if courses allow AI for homework, tests, etc, that is likely going to impact what those students learn and retain. I don’t have any answers, because in the big picture we are closer to all classes allowing at least limited to AI vs. none (IMO). And at schools where ultimately AI might not be an option…IMO again, that school will have fewer customers.
Here’s an announcement from UChicago re: AI availability to staff/faculty/students:
As far as I can tell, AI is useful to do more quickly than you can, things you already know how to do and where you can accurately review returns/outputs.
It’s like, nowadays, we just type 3587×2066 into a calculator and don’t calculate by hand BUT must absolutely, need to, learn multiplication tables and complete problem sets till we know how all 4 operations work : they’re necessary for our brains to build the reasoning behind the tap tapping on calculator keys + need to have mathematical automatisms (? Automatic thinking?) to do anything.
In other words, AI may be useful once students have mastered the knowledge, skills, and process - therefore AFTER college (or after a selected course content. Has been mastered) not DURING.
Yet students use AI to have it complete the process and avoid learning skills& knowledge they need. They unskill or deskill themselves by using it if it does the work they’re supposed to be learning rather than speed up the work they know how to do and can just check.
I was watching a Hollywood roundtable recently and I was actually shocked because on the one hand you had Karolyna Wydra and Michelle Pfeiffer on the side of nodding/explaining how if you “write” “with” AI, studies show you don’t remember what you’ve written, how using AI instead of your own brain makes you stupid, how dangerous that is, and on the other hand all other actors (I won’t give their names) laughing like it’s the biggest joke, unable to engage with the subject as a serious matter that requires thinking and not just deflecting/jumping on bandwagon - unable or unwilling to engage perhaps? They knew there were being filmed and that their reactions would be shown yet they behaved like brats specifically wrt this matter. That was truly shocking to me. (They were fine for the rest of the roundtable.)
Just to be clear I don’t support using AI on exams, but some profs/classes allow that. And some don’t, in the same school. Doesn’t seem to be tenable, or solid policy for a school to allow these extremes, was my only point.
I don’t know how AI is going to shake out for humans, but to me it seems different than some other developments in the past, like calculators and even computers. And the AI genie is not going back in the bottle.
I am not surprised that some/many people choose not to think, given that option.
I teach programming in high school. I have had several sincere discussions with my students regarding using AI to do assignments. My understanding is that a large portion of my students don’t truly understand why they study. In their mind they need decent grades so the adults won’t bother them and they could use their time to have fun. The fastest way to get the required grades is to use AI. Of 80 students last school year, I know 14 want to learn and genuinely do their work even if it takes time and are often flawed. That’s not even 20%. If the kids don’t understand why they should study, they won’t put in the effort to truly learn. I feel that many college students don’t understand the purpose of learning either, even those in highly selective colleges.
I think this is an excellent point (and has long been the case). Many students seem to think the purpose of studying is just to get that diploma at the end of it.
Currently, UCB EECS requires Math 54, but there were some years recently when linear algebra content was taught in EE 16A and 16B, and Math 54 was not required.
Other lower division linear algebra courses also exist (Math 56 and Physics 89).