Artificial intelligence has made its way into nearly every corner of education.
Students use it to research, write and study and teachers are still expected to determine whether the work in front of them is genuinely their student’s.
AI’s presence in education has raised numerous questions about academic integrity, but another concern deserves attention: who is responsible when the technology gets it wrong?
Students turn to tools such as ChatGPT and Gemini to research and complete coursework.
A University of British Columbia survey of 480 Arts students conducted between Fall 2024 and Spring 2025 found that 10 per cent said they use AI daily for coursework.
Students are increasingly held accountable for AI use but the companies that create AI are not being held to the same standard of accountability.
With a resource so readily available, distinguishing academic dishonesty has become more difficult.
lSome post-secondary institutions rely on educators and AI-detection tools to determine misconduct.
A 2026 Fraser Institute Leger poll study found that 42.2 per cent of Canadian teachers in grades six to 12 believe more than half of their students use AI to complete schoolwork.
While the figure focuses on grade-school teacher’s perceptions, it brings out the growing uncertainty educators could face when assessing student’s work.
AI detection software has its shortcomings. A 2023 study in the International Journal for Educational Integrity examined 14 detection systems and found that none consistently distinguished AI-generated writing from human-written work.
This leaves students to defend the authenticity of their work against a technology that is not always accurate.
A separate 2025 study by researchers Selin Dik, Osman Erdem and Mehmet Dik tested GPTZero with fifty human written works.
It found that GPTZero flagged human-written essays with an average AI-generated score of 29.86 per cent. The score rose to 35.56 per cent for shorter essays.
The study's sample was limited but its findings raise questions about the reliability of AI-detection tools that AI companies have yet to provide all the answers to.
Artificial intelligence companies must take greater responsibility for the technology they put into the hands of educators and students.
The detection companies should be required to regularly test their products and improve their accuracy.
They should establish safeguards for their products that ensure it is not treated as definitive proof of academic misconduct.
Some schools have started to move towards no AI in the classroom.
The University of Toronto banned computers in multiple classrooms on campus this year.
This is one way schools can step up and do what they can to prevent the AI from taking over classrooms.
