AI vs Teacher: How a hidden trick is revealing students who use ChatGPT, Google Bard in class

The advent of AI chatbots like ChatGPT and Google’s Bard has ushered in new concerns about academic integrity. These tools, designed to generate human-like text based on the prompts they receive, have become a tempting option for students looking to circumvent the traditional essay-writing process. This development poses a significant challenge for educators striving to maintain the integrity of academic assignments. But some teachers are finding their own little ways to solve this pressing problem.

Daina Petronis, an English teacher from Toronto, Canada has garnered attention for her innovative approach to identifying essays generated by AI. In a TikTok video cited by Daily Mail, the teacher has grabbed both interest and criticism. Petronis unveiled a method she refers to as a ‘Trojan Horse.’ This technique involves embedding a hidden instruction within the essay prompt that directs the inclusion of specific, unrelated words in the student’s submission. The instruction, formatted in white text and minimized in size, remains unseen by students but is detected by AI when the prompt is copied into a tool like ChatGPT.

The process Petronis outlines:
1. Divide the essay prompt into two parts. 2. Embed a directive in one part asking for the inclusion of arbitrary words. 3. Conceal this directive by setting the font color to white and reducing its size. 4. Reassemble the prompt.

When students use ChatGPT, the AI ​​incorporates the specified ‘Trojan Horse’ words into the essay, signaling to the teacher that the work may not be the student’s own. Petronis’s method relies on the premise that AI, unlike humans, will not overlook these hidden instructions and will dutifully integrate the specified terms into its response.

This approach also sparked discussions among educators and students alike. Critics argue that it primarily targets students who use AI without reviewing the generated text, suggesting it might not catch more diligent cheaters.

Petronis emphasizes that the primary goal of her ‘Trojan Horse’ technique is to ensure academic integrity and to provide an opportunity to guide students away from reliance on AI for educational tasks.

Addressing the surge of chatbots and generative AI, a Harvard Graduate School of Education study provides some guidance for educators in the era of AI. The report made the following points that could help both educators and students deal with the advent of the new technology:1. AI is Here: AI is becoming a significant part of our world, and it’s crucial for teachers to understand it. Even though it might feel unfamiliar or intimidating, it’s a reality we need to embrace.2. AI is a Tool: AI can serve as a beneficial tool in the classroom. It can assist teachers in understanding students better and provide students with a more personalized learning experience.3. Teachers Need to Learn About AI: It’s essential for teachers to learn about AI so they can guide their students. Students are already using AI tools, and they need to understand how to use them responsibly.4. Use AI With Students: Teachers should incorporate AI tools with their students. This can be done in class or assigned as homework. The objective is to help students understand how AI works and how to use it responsibly.5. Teach Students to Ask Questions: One of the most vital skills students can learn is how to ask good questions. This is something that AI can’t do, so it’s a skill that will always be valuable.