It Is Not AI That Threatens Education, But the Loss of the Thinking Process
In a lecture hall in Yogyakarta, a lecturer observed a significant shift. The use of artificial intelligence (AI) allows lecturers to complete the preparation of Semester Learning Plans, presentation materials, and exam instruments in just a few minutes. This productivity and work efficiency indeed has a tremendous impact. However, another reality among students is also becoming clearly visible. Course assignments are now submitted with much neater systematics, more organised writing structures, and seemingly more abundant reference lists. A fundamental problem arises when students must re-present or defend their arguments in discussion sessions, where some of them actually experience difficulty. This pattern indicates that the use of AI accelerates the acquisition of answers, but does not necessarily enhance the depth of students’ conceptual understanding.
This reality reflects a situation that is common in academic environments. The presence of generative AI technology in various universities has revolutionised the methods students use to review literature, compose scientific papers, and complete coursework demands. This fact proves that the skill of producing a text is now detached from the capacity to understand the substance of the material itself. Through instant processing, this technology can formulate essays, extract key points from scientific articles, and construct argument frameworks with seemingly valid structures. Therefore, the main challenge in higher education today has shifted; it is no longer about students’ ability to present solutions to an academic problem, but about the originality of the cognitive process underlying the birth of that answer.
This technological transformation demands serious attention because artificial intelligence has now become integrated into the education system. Various countries, including Indonesia, are drafting regulations and strategic policy directions regarding AI integration in the public sector, particularly within university environments. This fact confirms that AI is no longer a temporary supplementary instrument, but a fundamental element of the knowledge infrastructure firmly embedded in academic activity. This condition makes discussions about regulations permitting or banning AI use in educational institutions no longer relevant. The main challenge that the academic world must resolve is formulating the right mechanism so that the adoption of this technology does not degrade human critical thinking abilities.
From a cognitive psychology perspective, this situation represents the phenomenon of cognitive offloading, which is the tendency of individuals to transfer part of their cognitive load to external devices. This practice has actually been going on for a long time, such as using calculators for numerical calculations, search engines to replace retentive memory, and digital navigation systems to replace spatial orientation skills. However, generative AI has a much broader reach because this technology does not merely store or find data, but is capable of constructing, organising, and producing a thought comprehensively.
Some research indicates that the use of generative AI can optimise learning efficiency and facilitate the mastery of complex concepts. However, this series of studies also identifies the threat of reduced cognitive engagement if students adopt AI outputs directly without going through stages of verification and critical reasoning. In this condition, the position of AI has shifted from a reasoning support instrument to an absolute replacement for the thinking process itself. This condition contradicts the essence of higher education, which is not solely oriented towards achieving final results, but towards the formation of students’ thinking structures and methodologies.
Another crucial challenge lies in the inherent characteristics of artificial intelligence itself. Large Language Models can produce responses with highly convincing linguistic structures, but their substantial accuracy is not always guaranteed. The phenomenon of AI hallucination proves that the system can formulate incorrect information, fabricate fictitious reference lists, and even formulate conclusions without valid data support. If students do not possess strong critical reasoning, they will easily accept outputs that appear scientific as factual truth.
In various universities, the use of AI has now merged with daily academic activities. Students rely on this technology to find research topics, construct frameworks of thought, and refine their scientific writing style. On the other hand, lecturers also use it as a supporting instrument in preparing lecture materials and evaluation instruments. This adaptation process has become an unavoidable reality and has proven to contribute positively in many aspects. Nevertheless, this condition triggers new problems regarding the right mechanism to ensure that all this technological convenience does not eliminate the essence of the learning process itself.
Therefore, current educational strategies can no longer rely on policies of absolute prohibition or uncontrolled permissiveness. The urgent need now is a transformation in the methods of evaluating student learning outcomes. Assessments based solely on final products, such as essays or scientific papers, are no longer representative for measuring students’ original competence. The focus of assessment must be returned to tracing the process.