Generative AI should end the failed war on school misconduct

Generative AI should end the failed war on school misconduct

Higher educational institutions have long been waging a Sisyphean struggle with students who are ready to write off tests and assignments. Considering how generative AI reveals the limitations of current evaluation modes, Utkarsh Leo argues that educators and employers should consider how an obsessive focus on grades has skewed teaching and learning in universities.


Despite the long struggle to end it an increase in student academic misconduct poses a threat to higher education in the UK. Why is this so, and is there anything that can be done to fix it?

Misconduct in Higher Education (HE) Not new. Advances in technology, from the launch of the World Wide Web and the advent of Google to rogue wearable gadgets like the Monorean, have led to misconduct. comfortable. The possibility of deception has only increased with the advent of text-generating artificial intelligence models. This is evidenced by recent poll of 1,000 undergraduate students who found: 43% of student respondents admitted to using ChatGPT or similar AI applications; 30% admitted to using AI for most of their jobs; and 17% agreed to submit an AI-generated term paper without edits.

Rice. 1 Use of AI tools by college students, source: Top Colleges.

1689940728 889 Generative AI should end the failed war on school misconduct | lifefromnature

To what extent do you use artificial intelligence tools such as ChatGPT to complete your assignments, exams, or related school assignments? (N=216), Source: Top Colleges.

The unethical use of AI or any form of academic assessment fraud is a concern. This destroys the student learning process, reducing the opportunity to hone critical thinking, writing, and other academic skills. First of all, fraud harms society, as it produces professionals who lack the necessary skills.

As gatekeepers, higher education institutions (HEIs) work to protect academic integrity– from partnership with Turnitin (text matching software) to signing Charter of Academic Integrityto design university rules, code of conduct and policies. Despite this, their efforts unsuccessful. This is because the approach to punishing misconduct through disciplinary measures (downgrading, module failure, or exclusion) is of little use when the likelihood of detecting misconduct remains low. Moreover, “grades are treated as the currency of learning” and a stepping stone to better career opportunities. As a result, our policies and regulations remain inadequate to address the real incentives behind student cheating.

Students “conscious decision makers“. They decide what is good for them based on the options available. From an economic point of view, they choose the option that provides maximum utility. As the prime minister recently and controversially stated, students choose higher education because they expect beneficial benefits. (such as, increase in earningssocial status) exceeds expected costs (for example, to pay for commissions or opportunity costs of lost profits). Even today”obtaining a bachelor’s degree is financially beneficial for most students“. It’s confirmed data that show the number of postgraduate students increased from 271,475 in 2012-2013 to 493,045 in 2021-2022; while the number of first degree students increased from 495,325 to 648,925.

1689940729 567 Generative AI should end the failed war on school misconduct | lifefromnature

Fig. 3 Enrollment of students for the first year of study (HE) by levels of study, 2012/3–2021/22 academic years, source HE.

This raises the question: if students are rational and the higher education process brings higher returns; then academic misdeeds are irrational as they result in zero learning? Legal and economic analysis shows that, despite the lack of knowledge, fraud (in online/on-site exams) is the shortest path to good grades (higher classification), leading to broader career prospects. A review of job descriptions shows that a number of employers are often looking for graduates with at least 2:1 under their belt or those with ‘strong academic background‘. Similar incentives exist for those interested in postgraduate education.

Cheating incentives become more lucrative when the likelihood of being caught for academic misconduct is low. At most UK universities, the likelihood of detecting misconduct is low because assessors rely on Turnitin to check the originality of student papers. No matter what Turnitin’s accuracy and effectiveness are debatable. Turnitin himself emphasizes that he does not give an unambiguous judgment about plagiarism. Rather, his account of similarity support academic judgment appraiser.

Academic judgment, a major determinant in detecting misconduct, is impaired due to time limit and volume of student applications in the UK. Grading and providing detailed final feedback is a task that must be completed within the time frame for moderation, boarding, and publishing the results. Carrying out meaningful checks to identify misconduct and to impose proportionate fines (requiring hard evidence, as confirmed by Office of the Independent Judge) is rare. In this way, cheating remains profitable, as most cases of suspected plagiarism are considered due to poor academic practice, resulting in lower grades—a beneficial outcome for a student who might otherwise fail.

Unfortunately, a similar approach is being taken to combat misconduct with the help of AI. Quantity institutions implemented a new Turnitin with AI letter discovery capabilities (while others have expressed interest in take it later). However, since generative AI models are constantly evolving and improving when creating human-like content, it is unlikely that any AI letter detection tool can tag submitted materials as AI built with confidence. If a student (with English as a second language) chooses to use an AI paraphrase tool (such as Quillbot) to improve the quality of their written work, the AI ​​Detector will mark the submission as AI generated. This will only get more complicated as AI tools are introduced. built into MS Office. Of even greater concern is the possibility of GPT detectors that “exacerbate systemic biases against non-native authors“, mistaking their original work for AI-generated work. Therefore, appraisers are likely to shy away from implementing measures to avoid unfounded accusationswhich can lead to negative feedback from the regulator.

Unfortunately, “we have created a system where we value grades more than learning.The technology to allow cheating will improve and it will be difficult to play catch-up. Solving this problem will require joint efforts of both universities and employers. Our upcoming research shows that misconduct can be limited if universities require students (with a focus on international students) complete the required training for each year of study to instill the fundamental values ​​of academic integrity. Module leaders may consider focusing on experiential learning to improve student learning experience. For example, based on my own teaching experience, it may be more effective to teach modules such as “Alternative Dispute Resolution” (more specifically, hostile negotiation and decision making skills) on simulators that “realistically recreates (…) a crisis situation(e.g. the Hydra/Minerva kit at the University of California).

In addition, from an assessment standpoint, it may be beneficial if we change the way students are tested. For example, essay-style grades (which have promising potential for plagiarism) may be limited to one per semester in a module of interest to the student. Whereas, assessment for other modules may use a combination of work done on real cases (for law students, this could be a law school clinic), internships, project presentations, and time-limited problem-style physical exams. From a skills development point of view, it would be beneficial if participation and excellence in extra-curricular activities (eg Moot Court) could result in a submission/score for the competition being withdrawn. First of all, to improve academic judgment, universities could consider rewarding good grades. For example, the quality of supervision and the resulting feedback may be taken into account by the line manager for promotion.

At the other end of the pipeline, it is critical that potential employers take action to combat misconduct, as those who commit academic misconduct are more likely to engage in dishonesty in the workplace. Therefore, should hiring be based on grades? No. Applicants must be tested on the skills and application of knowledge to solve real problems.

All in all, if we approach the problem in the right way, we can succeed in limiting wrongdoing and de-escalating the obsession with grades in favor of learning, which is, after all, the true purpose of education.


I am grateful to Sean Mills, Associate Dean of the UCLA School of Justice, for his invaluable comments and feedback.

The content created on this blog is for informational purposes only. This article represents the views and opinions of the authors, but does not reflect the views and opinions of the Impact of Social Science blog (blog) or the London School of Economics and Political Science. Please see our comment policy if you have any concerns about posting a comment below.

Image credit: Sergey Zolkin via Unsplash.com.


Printable, PDF and Email