Post

Conversation

There is a field experiment showing this exact effect. Introducing GPT tutors increases performance by *a lot*--students seem to be picking up the material much faster--but when GPT is removed those who had access perform *much worse* compared to those w/o access. 1/4

Quote

Lakshya Jain

@lxeagle17

Mar 12

I'm teaching databases this semester at Berkeley. My students all seem unusually brilliant. Not many go to office hours, and not too many folks post on the course forum asking project questions. Weirdly, the exam had the lowest recorded average in my 10 semesters teaching it.

8:56 AM · Mar 13, 2025

689.3K

Views

Post your reply

Alex Imas

@alexolegimas

Mar 13

Brings question: what do we want students to learn in our classes? There is an argument that students will always have access to GPT, so assignments should allow access. But another view is that the class is not just teaching material, but how to *think* through problems. 2/4

Each field's questions have logic to them--field-specific critical thinking skills--that goes beyond specific questions. Using LLMs will answer question but students will fail to learn how to think through q's. This will lead to disadvantage even when LLM available. Why? 3/4

There will be questions that the LLM will get wrong, and student will not be able to diagnose why or that it's wrong in first place. Perhaps more importantly, student will have less capacity for creativity because they don't understand the logic behind the field's questions. 4/5

This is why I've changed my classes to focus more on in-class assignments. It's not ideal--big final projects are fun and important, and they're still a component of the class--but it's also important that students are leaving class knowing the logic of how to think through qs.

Reminds me of this RCT where they gave incentives to kids to improve math skills, then removed the incentives. Good students remained good, bad students ended up providing lower effort in every task and performed worse than before.

Very relevant!

The 2x Ether ETF (Ticker: ETHU) is a leveraged ether-linked ETF. ETHU seeks to benefit from increases in the price of Ether. Learn more about ETHU today. ETHU invests in ether futures and does not invest directly in spot ether.

ETHU ETF: Leveraged Ether-Linked ETF

From volatilityshares.com

I think a major issue with moving to in-class assignments, is that it's hard to teach long-form writing that way. Essays require multiple revisions and research over the course of days.

Absolutely. Unless this issue is solved through smart implementation of the tech, this will be an important skill that could become underdeveloped.

I like this study and cite it but I have two points of why this isn’t settling the issue of AI and learning 1) it’s a field experiment on a topic where its impossible to measure treatment since there’s no AI detection tools since AI use can’t be reliably detected

But doesn’t that mean it’s a conservative test of harm on learning?

Do you have a link to this paper?

4.4K

Alex Imas

@alexolegimas

To view keyboard shortcuts, press question markView keyboard shortcuts

Post

Conversation

To view keyboard shortcuts, press question mark
View keyboard shortcuts