Opinion
Artificial Intelligence Opinion

Three Steps to Prevent ChatGPT Misuse

Original or AI? How teachers can prevent plagiarism
By Spencer Burrows — March 07, 2023 4 min read
Conceptual vector illustration of Women in AI examining a virtual apple.
  • Save to favorites
  • Print

The remarkable technology known as ChatGPT has recently taken social media and the chattering class by storm. By now, many have witnessed at least one incredible product created by the artificial intelligence bot, be it a technical paper on coffee bean cultivation or a song about earthquakes in the style of Taylor Swift.

However, many teachers quickly realized a downside to the amazing technology: students potentially using the app to plagiarize assignments. First remote learning, then hybrid learning, and now this?

Yet, there are steps teachers can take to prevent the misuse. First, it is important to understand what ChatGPT is and is not. ChatGPT is a word-prediction technology, not an independent wealth of knowledge. That means the program needs to make a connection between real subjects to generate a fair word prediction. The app can draft a reasonable sounding paragraph on the link between “Aristotle” and “democracy” but not on a “thing” and “democracy.”

See Also

Illustration of field of word bubbles with technological elements superimposed.
Vanessa Solis/EdWeek + iStock/Getty images

While ChatGPT is not necessarily producing lies (intentional untruths), it is very often producing baloney. How to prevent this plain nonsense from infiltrating schoolwork? It all comes down to how you phrase your assignment questions.

Here’s what I recommend:

1. Avoid identifying terms and instead use signpost terms.

As I mentioned, ChatGPT needs real terms to produce a decent sounding response. I currently teach high school U.S. history and assigned a reading that covered Thomas Jefferson, the Declaration of Independence, and the Revolutionary War. Had I phrased the assignment question as: “How did Thomas Jefferson write the Declaration of Independence?” The ChatGPT response is:

“Thomas Jefferson wrote the Declaration of Independence by drafting an outline of grievances against the British government and a call for the colonies to declare their independence. He then wrote a rough draft of the document that was edited and revised by Congress over a period of several weeks. Jefferson’s final draft of the document was officially released on July 4th, 1776.”

That is a fair sounding response from ChatGPT.

However, look at the answer when I phrased my question as, “How did the author on Page 4 write that document?” The ChatGPT response to this second question was:

“The author on page 4 wrote the document using a combination of formal and informal language. They used a structure that is clear and easy to read, while also making sure to include specific details and relevant facts related to their subject. The author incorporates both personal opinion and open-mindedness in their writing, while also relying on research and data to back up their points.”

You can see here that ChatGPT has no idea who the author is and produced a baloney response. Hence, I encourage teachers to frame your assignment with language that requires the students to go back to the text and find the answer, such as using specific page numbers or references.

2. Require text citations in the response.

After asking the students to return to the text to find the answer or produce the analysis, ask for text citations as well. This really sends ChatGPT on a fishing expedition.

I asked the app, “Write one paragraph explaining how the author on Page 4 helped win the war. Use at least two citations from the text.” The response:

“The author on page four describes how Joseph Pilsudski, the founder of modern Poland, had a major role in helping win the war. Pilsudski mobilized a Polish army and forced the Austrians to sign an armistice, formally ending their occupation of Poland in 1918 (pg.4). He also formed a multi-national Polish army and ‘fought shoulder to shoulder with his former enemies against the Bolshevik menace’ (pg.4), helping Greece, Romania and Czechoslovakia gain independence from the Bolsheviks. Without Pilsudski’s efforts, the war with the Bolsheviks would have been much more difficult and prolonged and the outcome of the war would have been far different.”

Complete nonsense. ChatGPT has no idea who the author is and even less idea what war we are referring to, so students cannot effectively plagiarize an assignment that requires citations.

3. Ask for a modern-day connection in the analysis.

ChatGPT is not nearly as adept at generalizing on new or developing situations. When the app is asked: “How is the Revolutionary War related to the crisis in Ukraine?” The response is:

“The Revolutionary War is not directly related to the crisis in Ukraine. However, certain parallels can be drawn between the motivations of the revolutionaries in the American Revolution and those initiating the revolution in Ukraine. In both cases, people have sought a more equitable distribution of political and economic power, a greater say in governance, and more freedom from external control.”

Again, boilerplate language. The app took “revolution” and generalized language that could have described the Russian Revolution, the Cuban Revolution, or most any revolution for that matter. Requiring students to apply the subject to current events, therefore, lowers the likelihood they will outsource their thinking to AI.

I hope these tips keep you afloat until spring break. Best of luck!

Events

This content is provided by our sponsor. It is not written by and does not necessarily reflect the views of Education Week's editorial staff.
Sponsor
Artificial Intelligence Webinar
From Passive to Active Reading: Student Engagement in the Age of AI
What the research tells us about active reading and student engagement—and what it means for teaching in an AI-enabled world.
Content provided by Pearson
This content is provided by our sponsor. It is not written by and does not necessarily reflect the views of Education Week's editorial staff.
Sponsor
Professional Development Webinar
Micro-Learning, Macro Results: Reimagining Professional Learning
Rethink professional learning with 3 strategies that personalize professional learning, validate expertise, and support educator retention.
Content provided by Mira Education & Modiv EDU
This content is provided by our sponsor. It is not written by and does not necessarily reflect the views of Education Week's editorial staff.
Sponsor
School Climate & Safety Webinar
There But Not Present: Rethinking Student Engagement
Learn how to recognize the early signs of disengagement and what the behavioral science reveals about which interventions actually work.
Content provided by Frontline Education

EdWeek Top School Jobs

Teacher Jobs
Search over ten thousand teaching jobs nationwide — elementary, middle, high school and more.
View Jobs
Principal Jobs
Find hundreds of jobs for principals, assistant principals, and other school leadership roles.
View Jobs
Administrator Jobs
Over a thousand district-level jobs: superintendents, directors, more.
View Jobs
Support Staff Jobs
Search thousands of jobs, from paraprofessionals to counselors and more.
View Jobs

Read Next

Artificial Intelligence Q&A An Educator Banned Social Media in Her School Years Ago. What About Banning AI?
An AI safety expert and former educator spoke at a UN meeting about AI safety. This is what she said.
5 min read
Dario Amodei, CEO of Anthropic, speaks remotely during a Security Council meeting on artificial intelligence at the 81st session of the United Nations General Assembly at United Nations headquarters, Wednesday, Sept. 23, 2026.
Dario Amodei, CEO of Anthropic, speaks remotely during a Security Council meeting on artificial intelligence at the 81st session of the United Nations General Assembly at U.N. headquarters on Sept. 23, 2026. Another meeting that took place at the General Assembly focused on AI safety for children. Erin Mote, a former school leader, participated and discussed the importance of tackling the challenge globally.
Yuki Iwamura/AP
Artificial Intelligence Opinion Why I'm Not Betting on the Future of Alpha School
A former champion of the ed-tech movement shares his concerns about this new school model.
4 min read
Conceptual illustration of classroom conversations and fragmented education elements coming together to form a cohesive picture of a book of classroom knowledge.
Sonia Pulido for Education Week
Artificial Intelligence Headlines Warn of an AI Apocalypse. How Should Schools Respond?
Panic-inducing AI headlines might prompt schools to hit pause and review the tech more carefully.
6 min read
Scary Headline AI Illustration
Vanessa Solis/Education Week + Getty
Artificial Intelligence Can Schools Trust AI to Translate for Non-English-Speaking Parents?
AI tools can translate school materials for non-English-speaking parents, but there are challenges around privacy and accuracy.
10 min read
Illustration by Nicole Xu showing a teacher in a teacher-parent conference with a laptop that translates a distorted depiction of the student. The female parent of the EL student has a swirly white line of confusion circling her head.
Nicole Xu for Education Week