Universities Revamp Assessment Practices as AI Use Grows

Colleges across the United States and elsewhere are confronting widespread student use of generative AI tools such as ChatGPT. Surveys show that 73 % of faculty have encountered AI‑related integrity violations and major campuses report majorities of students completing assignments with AI. In response, many institutions have abandoned AI‑detection software, citing false‑positive and bias concerns, and are shifting toward redesigning assessments—emphasizing process, oral exams, proctored tests, and documentation of AI assistance.

Research from Cardiff Metropolitan University highlights fundamental weaknesses in large language models, noting their lack of human‑like socialisation and moral judgment, which can lead to confident but inaccurate outputs. Studies of student learning show mixed effects: while AI can boost homework scores, it may depress performance on exams without AI, and over‑reliance can erode critical thinking. Surveys at American University’s Kogod School of Business reveal that AI use among business students has risen sharply, with demand for formal AI‑literacy training.

The World Bank’s 2026 Development Report warns that AI‑driven job losses will hit countries such as Pakistan hardest, underscoring the broader economic stakes of AI integration in education. Educators and researchers call for a “pedagogy of kindness” that embeds supportive structures rather than relying on individual goodwill, aiming to ensure AI serves learning rather than undermines it.