Cheating, and AI Detectors

CTE - AI Assessments and ai

Cheating and Ai detectors

Finger pressing a laptop keyboard.

In this section,  Cheating and AI Detectors, we'll explore how students cheat, how AI detectors work, and whether or not we should use them. 

Are students cheating more?

The increasing presence of AI in education has sparked growing concerns about potential academic dishonesty and cheating in schools. However, it's essential to critically examine whether these concerns align with the actual trends and behaviors observed among students. By doing so, one can gain a more accurate understanding of the current landscape and devise effective strategies to address genuine issues related to academic integrity in the context of advancing technology.

Stanford researchers Links to an external site. Denise Pope and Victor Lee challenge the widespread concern that the advent of AI chatbots, including ChatGPT, is facilitating increased cheating among students. Their ongoing research into U.S. high school students' cheating behaviors reveals that, contrary to fears, there is no significant rise in cheating rates associated with AI tools.

The researchers emphasize that cheating is a long-standing issue and argue that the focus should be on addressing its root causes, such as academic pressure and a lack of engagement, rather than solely targeting AI technologies. They suggest fostering discussions about the ethical use of AI in education and promoting AI literacy among students, emphasizing responsible usage of technology.

a student on her computer in the mountains

Image created using AI

 

Should we use AI checkers?

As the integration of artificial intelligence continues to reshape the educational landscape, a pressing question emerges: Should schools embrace AI checkers to monitor and prevent academic dishonesty? Exploring both the potential benefits and ethical concerns, we weigh the advantages of technological assistance against the need to foster a culture of academic integrity and responsible technology usage.

magnifying lens on a computer

Image created using AI

How do AI detectors work?

AI detectors operate through sophisticated computer systems designed to recognize and categorize diverse objects or occurrences. These systems leverage artificial intelligence algorithms to meticulously analyze extensive datasets, enabling them to make precise predictions. Widely embraced across sectors like healthcare, security, and manufacturing, AI detectors offer efficiency in various tasks.

However, concerns persist regarding their accuracy. Despite notable advancements, instances of inaccurate detections or false positives raise questions about the reliability of these systems. This challenge is attributed to limitations in the underlying algorithms and inadequacies in training data. Researchers and developers are actively engaged in ongoing efforts to enhance the precision of AI detectors, aiming to address these concerns and optimize their performance.

Are AI detectors accurate?

Many "companies have developed “AI detection” software. This software aims to flag AI-generated content in student work. However, early experiences show that AI detection software is far from foolproof—in fact, it has high error rates and can lead instructors to falsely accuse students of misconduct (Edwards, 2023; Fowler, 2023). OpenAI, the company behind ChatGPT, even shut down their own AI detection software because of its poor accuracy (Nelson, 2023)" (MIT 2023 Links to an external site.). 

Given the concerns surrounding the reliability of AI detectors, with instances of inaccuracies and potential harm, a more prudent approach is to advocate for the development and implementation of meaningful and authentic assessments. Relying solely on AI detectors for academic integrity poses risks, as their limitations may lead to both false positives and negatives, impacting students unfairly. Authentic assessments, on the other hand, involve tasks that reflect real-world scenarios, requiring students to apply their knowledge and skills in relevant contexts. By prioritizing such assessments, educational institutions can foster a more comprehensive understanding of students' capabilities while minimizing the reliance on potentially flawed technological solutions. We will give you some ideas on the next page. 

a drawing of a computer on a desk

Image created using AI

Further resources

Are students really cheating with AI? https://ed.stanford.edu/news/what-do-ai-chatbots-really-mean-students-and-cheating Links to an external site.

International Students are falsely accused of cheating: https://themarkup.org/machine-learning/2023/08/14/ai-detection-tools-falsely-accuse-international-students-of-cheating Links to an external site.

AI detectors are biased against ESL Students: https://www.cell.com/patterns/fulltext/S2666-3899(23)00130-7?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS2666389923001307%3Fshowall%3Dtrue Links to an external site. 

Ramlochan, S. (2023, November 6). The truth about AI detectors - more harm than good. Prompt Engineering. https://promptengineering.org/the-truth-about-ai-detectors-more-harm-than-good/ Links to an external site.