9:58

supplementary video · PBS NewsHour on YouTube ↗

arts, ideas & societyanalysisThe Conversation (Australia)

AI has turned university assessments into a ‘wicked problem’ – and we don’t have all the answers

the gist

Since ChatGPT arrived in late 2022, universities have struggled to tell whether an assessment tests a student's learning or just their skill with AI. A Turnitin report found nearly 54% of Australian student submissions processed between October 2025 and April 2026 showed some AI use, and about 10% were more than 80% AI-generated. But AI use isn't automatically cheating: students use it to summarise readings, clarify hard texts, brainstorm and get feedback. The line between acceptable and unacceptable use is blurry, even to teachers, and can differ between assignments and even between students on the same one. Australian universities must still show that degrees confirm the skills they claim.

so what

This is what policy researchers call a wicked problem: no single fix, and every rule shifts the problem elsewhere. How universities redesign exams will decide what a degree is worth in an age of AI.

fun fact

Turnitin is best known for plagiarism-detection software used by schools and universities.