If I get the sense enough are cheating with stuff like this, we are moving a substantial portion of the final grade into the stuff we do in person in the class room
Comments
Log in with your Bluesky account to leave a comment
As a former Cincy State instructor: THANK YOU. I worked my ass off to get students prepared to get to your class ready to do the damn work, this shit is like a smack in the face.
I'm pretty bullish on LLMs being helpful tools for a whole bunch of applications, but I tend to think that a "no AI in class" position is justifiable. That said, I am curious to see how people actually justify it, in practice.
1. Students need both skills and domain knowledge.
2. LLMs are not great at history and humanities.
3. LLMs are built on stolen intellectual property.
4. Power and cooling requirements are indefensible.
5. Labor issues related to training LLMs.
I think I'd have nitpicks or just fundamental disagreements about 3-5, but 1 & 2 are indisputable and seem sufficient to justify the policy. I do think the calculus would be completely different for a discipline like software development or business, though.
The thing is for most undergraduate classes, especially in the humanities, the process of completing the assignments is more important than the actual product. Students need to do the intellectual work of learning the material, thinking about how it relates, and working with primary sources.
Comments
1. Students need both skills and domain knowledge.
2. LLMs are not great at history and humanities.
3. LLMs are built on stolen intellectual property.
4. Power and cooling requirements are indefensible.
5. Labor issues related to training LLMs.