What to Do About AI in Colleges?
There is a growing perception among students that ChatGPT can answer any
question they have. This feeling extends to young professionals as well.
Faculty must address this significant issue, perhaps even more urgently than
traditional political problems. Let’s explore this further.
Mathematics
With the advent of tools like Maple and Mathematica*, students quickly
realized these tools could solve many well-defined problems. These specialized
AI tools are continually improving. Students understand this but accept that
they must learn the techniques to pass their courses.
Engineering
Until recently, AI did not perform well on some engineering exams, but
its capabilities are improving. Despite this, students believe AI will be
available when needed in the future. As with math classes, students still
comply and learn the material.
Practical Application
What students often fail to see is that in practice, problems are not
simply presented; they must be created from underlying situations with
contingencies. Students need to see examples of this by working through complex
situations where they must formulate a problem-solving strategy. AI does not
yet do this.
Writing
At many colleges, students are not required to take a composition course;
instead, writing courses are taught according to department-specific criteria.
This is unfortunate. However, AI can assist here. For example, you can supply a
draft of your written work to ChatGPT and ask it to rewrite it. You might find
it smooths out your writing, cleans up the grammar, and reorganizes the
material better. I regularly use this feature, and it is helpful. These Large
Language Models (LLMs) are totally impressive.
Softer Sciences
In the softer sciences, students are often assigned topics or papers.
What’s to prevent students from simply asking AI to compose a paper on the
topic? AI has such a vast information base that it often generates papers
beyond the scope of most students. The question is: Why shouldn’t students use
these AI tools, at least to get started? The significant disadvantage is that
they never learn to assemble disparate information sources into a coherent and
consistent product. They never learn where to look or how to gather
information. What can be done? Students must now be confronted with situations
where they must create a problem to be solved and then write about it. This
should become the principal part of the assignment. AI cannot do this — yet (I
think).
Verification
You cannot check the output for AI intervention. For example, one can
easily switch between AI tools, asking for multiple rewrites. This makes it
difficult to verify originality.
This is just a brief introduction to the problems AI brings to learning,
and they may not be receiving the attention they need. AI is changing
everything. Students also need to learn what AI cannot do and base their
lifelong learning on this new reality.
*Maple and Mathematica are specialized software tools that do symbolic
mathematics, statistics, and high-level programming. They are essential tools
for many scientists and mathematicians.
P.S. This version is the AI-generated rewrite of my original rough draft.
----------------
© 2024 G Donald Allen
Comments
Post a Comment
Please Comment.