A neural network is a simulator, not a teacher: why critical thinking depends on pedagogy, and not on neural networks Natalia Karamzina, vice rector for educational and methodological activities at Vernadsky KFU, said that the problem of using artificial intelligence in universities is not related to the neural networks themselves, but to institutional delay: students are already working en masse in the new technological reality, but the evaluation system remains the same. "It's not just about the neural network, it's about the gap between training tools and control tools," Karamzina said. She emphasized that it is important to distinguish between using AI as a tool and replacing intellectual effort with a ready-made answer. In her opinion, there is no universal detector, so a verbal defense of the progress of the work and a transparent statement about the use of neural networks are more reliable. She added that artificial intelligence helps to develop critical thinking only with properly structured pedagogy - in itself it is a simulator, not a teacher. According to Karamzina, universities need local regulations that do not prohibit artificial intelligence, but define acceptable scenarios for each type of task. Also important are neural network-resistant assignments—case studies with fuzzy data, mid-course defense projects, oral colloquia—and teaching teachers how to integrate technology into the classroom. Image created by AI Listen to the broadcast of "Radio Crimea".

A neural network is a simulator, not a teacher: why critical thinking depends on pedagogy, and not on neural networks Natalia Karamzina, vice rector for educational and methodological activities at Vernadsky KFU, said that the problem of using artificial intelligence in universities is not related to the neural networks themselves, but to institutional delay: students are already working en masse in the new technological reality, but the evaluation system remains the same. "It's not just about the neural network, it's about the gap between training tools and control tools," Karamzina said. She emphasized that it is important to distinguish between using AI as a tool and replacing intellectual effort with a ready-made answer. In her opinion, there is no universal detector, so a verbal defense of the progress of the work and a transparent statement about the use of neural networks are more reliable. She added that artificial intelligence helps to develop critical thinking only with properly structured pedagogy - in itself it is a simulator, not a teacher. According to Karamzina, universities need local regulations that do not prohibit artificial intelligence, but define acceptable scenarios for each type of task. Also important are neural network-resistant assignments—case studies with fuzzy data, mid-course defense projects, oral colloquia—and teaching teachers how to integrate technology into the classroom. Image created by AI Listen to the broadcast of "Radio Crimea".