How Lincoln Public Schools prevents ChatGPT cheating
LINCOLN, Neb. (KLKN) – Technology has reached the point where computers can easily mimic human writing patterns.
Within seconds, the chatbot ChatGPT can turn out an essay on just about any topic you want.
It was released last November and quickly took on a life of its own among students nationwide.
The first-of-its-kind chatbot has the ability to write in a tone so humanlike, it could easily be passed off as a student’s work, if you don’t know what to look out for.
“They’re passable essays that got churned out, so we were like OK, we need to dig in more,” LPS curriculum specialist Sara Danielson said.
And that’s exactly what Lincoln Public Schools did when it banned ChatGPT from the servers that students use.
The district said it is determined to graduate students with the skills necessary to analyze objectively.
“We want our students to be able to think critically about the ideas that their having and how to move it to just this germ of a piece through the writing process,” Danielson said. “So, we want to make sure that we are working with students on how to think through and consider multiple perspectives and gather evidence, do those pieces that show and are visible evidence of thinking, not just, ‘I got an essay done.'”
Next LPS needed to find out how ChatGPT works and the thinking behind the program so it can detect when students try to pull a fast one.
After meeting with staff in the curriculum and computing departments, the district put out a guidance document in December that gave teachers examples of assignments generated by artificial intelligence so they have an idea of what to look for.
It turns out, it’s not as easy as some kids may think to get one over on your teacher.
“Teachers just know when it’s not their student’s writing,” Danielson said. “All of the sudden, it looks grammatically perfect, and everything leading up to it had all those typical developmental mistakes that we’re trying to work on, and all of the sudden, they just vanish.
Beyond that, Danielson said ChatGPT follows a basic template that’s easy to detect once you know it. For instance, it will use four-sentence paragraphs and the same transitional phrases throughout the essay.
If a teacher suspects a student of trying to cheat, they can plug the essay into detection software that will provide a percentage rate of the likelihood that the essay was generated through AI.
A handful of LPS students have tried using a chat bot to cheat. And LPS does have a dishonesty policy, so when a student is caught cheating, they may be expelled.
But LPS said artificial intelligence is not all bad and is willing to assist students who aren’t looking to cheat, but to study.
For example, a student might ask ChatGPT for a summary of a chapter as a study aid.
“There are a lot of applications for it,” Danielson said. “We just don’t want it to replace genuine thinking, you know, genuine authentic learning.”