A university lecturer is replacing coding tests with conversations about AI judgment

In an account published by Google on October 5, 2026, a lecturer and chief digital officer in Universiti Putra Malaysia's Faculty of Computer Science and Information Technology describes changing how data-mining students are assessed. Students use Gemini in Google Colab as a tutor that can suggest code and alternatives, but they must decide what to use, question the suggestions and justify their final modeling decisions.

The explanation becomes the test

Because an AI system can produce a working model, the lecturer focuses on whether students can exercise judgment over that output. In verbal interviews, students defend their choices, troubleshoot their work and identify recommendations from the AI that may be incorrect.

Gemini Canvas and Gemini Notebook help the lecturer evaluate the work across the class. The stated aim is to build lasting human capability while avoiding dependence on the technology, making the student's reasoning—not merely the generated code—the central evidence of learning.