College professors are searching for ways to “ChatGPT-proof” their exams
businessinsider.com
businessinsider.com
It's the education system that needs to adapt, not the students. Change is hard, but eras of computing drive societal change and professors and colleges need to adapt to stay relevant and not the other way around.
My favorite engineering professor was all over this decades in advance. You could bring any reference material you wanted to take one of his tests.
We couldn't find the solutions to his test problems in a search engine. If you can nowadays, maybe student and professor both need to reconsider your career path because you may not have one for very long.
Unless you are a contestant on Jeopardy, being able to recall lots of little pieces of data no longer has a lot of inherent value.
What does still have significant value is understanding what the data represents and how it can used and manipulated and combined and applied in creative ways in order to make good decisions and solve problems.
If you can't figure out how to describe your question, for example, because you haven't memorized enough vocabulary to express it, it becomes problematic.
Similarly if you aren't familiar enough with the topic to verify that the output of any such tool is reasonable, and you end up posting "rm -rf /" as an answer to a question of "how do I free space on my hard drive" because it's a common shitpost answer to the question.
And a lot of people have gotten through the education system, in my experience, by rote regurgitation from memorization without deeper understanding, so I think the changes around this are going to make much clearer to some a gaping divide between how well people learn right now versus how well people think others are learning.
The only way forward is performative knowledge verification in person and a controlled environment.
the running joke at university doing engineering was "here I have got two hours to get four problems half right, in the real world I am going to have as long as it takes to not make a mistake"
so, assess by continual assessment, make submissions by github, students have to push a commit at least every 20% of progress so the lecturer can see the incremental development/thinking and progressto wards the final result.
consequently, the correction of the approach and process to the problem can be much better tailored and it will also be much harder to use chatGPT, or any other cheating method for that matter.