By Rajkamal Rao
High school students already know that the most crucial subjective part of their college application is, without a doubt, the college essay.
Using A.I. chatbots to help write essays
Using chatbots like ChatGPT and Gemini is a relatively easy way to get to your first draft. It's free and quick and the grammar is excellent. You could modify the generated text to fit your personality. Many of our clients are embracing this approach when they submit their essays for reviews - and we can tell relatively easily because we know the quality of their writing from their prior writing engagements with us.
As the New York Times noted on September 1, 2023, A.I. chatbots are good at generating broad responses to short answer prompts. But our experience has been that chatbots don't do a good job at the longer platform essays, like the Common App Personal Statement.
According to the Times, colleges are conflicted about using chatbots, which "could facilitate plagiarism on college applications or democratize student access to writing help. Or maybe both." We agree, especially on the plagiarism front. We caught many more essays during the 2023 season where essays were blatantly plagiarized. We cautioned some clients with a generic warning: This is dishonest and could lead to your application being automatically rejected. Please redo this entire essay on your own and resubmit it.
Yale admission officers have reacted to the technology, even publishing a podcast on the do's and dont's: "A.I.-generated content simply isn't very good at the mode of communication that works in college essays."
With the increased use of chatbots, tools like Grammarly Premium (GP) have introduced an AI detection feature. GP warns writers about how much content was obtained using AI tools and even identifies a metric, such as "31% of the content was developed using AI."
But GP doesn't tell us how to fix AI detection issues - and in our tests, even if we don't use AI, it shows that we did. Even GP acknowledges this fact: "While AI detectors may play a role in investigating potential AI use, they are unreliable and can produce false positives on any written text, whether or not Grammarly was used." As large language models become bigger, someone, somewhere, has used similar content - and Grammarly flags it as AI detected. Here is another Grammarly post about AI detection and how it should be viewed.
Grammarly, however, does provide a reliable plagiarism score in the final report. We work with words, phrases, and sentences for each infraction until the plagiarism score comes up as clean.
Colleges have clear guidelines about plagiarism (an absolute no-no) but not about AI given its rapid advancement. We will adhere to this standard and flag anything that fails the GP plagiarism test. Just because GP suspects the use of AI and generates a new metric (probably to sell more licenses) doesn't mean much. It certainly has no material impact on how we review our clients' work or on college admissions. So, we ignore AI detection flags.
Colleges have never outlawed the use of chatbots in essays, such as to generate ideas and for research, as long as the information finally presented is original and does not involve plagiarism. We will adhere to this standard and flag anything that fails the GP plagiarism test.
But GP doesn't tell us how to fix AI detection issues - and in our tests, even if we don't use AI, it shows that we did. Even GP acknowledges this fact: "While AI detectors may play a role in investigating potential AI use, they are unreliable and can produce false positives on any written text, whether or not Grammarly was used." As large language models become bigger, someone, somewhere, has used similar content - and Grammarly flags it as AI detected. Here is another Grammarly post about AI detection and how it should be viewed.
Grammarly, however, does provide a reliable plagiarism score in the final report. We work with words, phrases, and sentences for each infraction until the plagiarism score comes up as clean.
Colleges have clear guidelines about plagiarism (an absolute no-no) but not about AI given its rapid advancement. We will adhere to this standard and flag anything that fails the GP plagiarism test. Just because GP suspects the use of AI and generates a new metric (probably to sell more licenses) doesn't mean much. It certainly has no material impact on how we review our clients' work or on college admissions. So, we ignore AI detection flags.
Colleges have never outlawed the use of chatbots in essays, such as to generate ideas and for research, as long as the information finally presented is original and does not involve plagiarism. We will adhere to this standard and flag anything that fails the GP plagiarism test.
College essays differentiate you from everyone else because when done well, they are personal narratives that are largely reflective of your experiences and introduce you to college admissions officers. A chatbot is a poor substitute here. The essay appears to come out looking good, but if it does not represent who you are, what good is it? Is it worth sacrificing your entire application because you didn't take the time or effort to invest in writing your essays?
Please contact us for more information.
A Note About Rao Advisors Premium Services
Our promise is to empower you with high-quality, ethical, and free advice via this website. But parents and students often ask us if they can engage with us for individual counseling sessions.
Individual counseling is part of the Premium Offering of Rao Advisors. Please contact us for more information.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.