自 ChatGPT 引入以来,澳大利亚大学在应对学生使用生成式 AI 进行课程评估方面面临挑战 1。Turnitin 数据显示,2025 年 10 月至 2026 年 4 月间,近 54% 的澳大利亚大学生提交物显示存在某种形式的 AI 使用 1。其中,约 10% 的提交物包含超过 80% 由 AI 生成的材料 1。研究指出,AI 与评估是一个“棘手问题”(wicked problem),没有单一的解决方案 1。
为应对这一挑战,悉尼大学采用了“双车道方法”(two-lane approach) 1。第一车道为“安全”评估(如线下考试),禁止使用 AI;第二车道为“开放”评估(如在线测验),允许自由使用 AI 并侧重于反馈和学习 1。然而,该领域仍被视为一个复杂的“棘手问题”,受限于成本、公平性及可穿戴 AI 设备等新技术 1。此外,学生已在新西兰、台湾和韩国被发现在考试中使用 AI 智能眼镜等可穿戴设备作弊 1。
Australian universities are grappling with the pervasive impact of generative artificial intelligence on academic integrity, facing what experts describe as a complex challenge without a single definitive solution. 1 Data from Turnitin reveals that between October 2025 and April 2026, nearly 54% of submissions from Australian university students showed evidence of some form of AI usage. 1 Within this group, approximately 10% of the submitted materials consisted of more than 80% AI-generated content, highlighting the scale of reliance on these tools. 1
To address this issue, institutions like the University of Sydney have implemented a "two-lane approach" to assessment design. 1 This strategy divides evaluations into two categories: a "safe" lane for assessments such as offline exams where AI use is strictly prohibited, and an "open" lane for tasks like online quizzes where AI use is permitted to focus on feedback and learning outcomes. 1 Despite these reforms, the situation remains complicated by factors including cost, equity concerns, and the emergence of new technologies. 1 Notably, incidents have already been reported in New Zealand, Taiwan, and South Korea involving students using wearable devices, such as AI-powered smart glasses, to cheat during examinations. 1
评论
还没有评论,欢迎留下第一条。