Page 245 - Izobraževanje v dobi generativne umetne inteligence

Učinkovitost generativne umetne inteligence za personalizirano učenje matematike

The Effectiveness of Generative Artificial Intelligence for Personalized
Mathematics Learning
Generative artificial intelligence (GEN-AI) is becoming an indispensable tool in ed-
ucation and mathematics instruction, as it enables the adaptation of tasks to the
knowledge level of individual students, thereby promoting personalized learning
and improving academic outcomes. Additional benefits include enhanced vocabu-
lary, increased curiosity, strengthened connection-building skills, improved ability to
reformulate questions, and deeper critical evaluation of answers. However, tasks gen-
erated by AI often fail to reflect the cultural and personal characteristics of students,
which can limit their authenticity. Research indicates that generative AI is particularly
beneficial for students with lower prior knowledge, as it facilitates faster progress. Its
greatest value is achieved when combined with active teacher support, who guides
the learning process and evaluates outcomes—a method known as hybrid tutoring.
In educational settings, generative AI can assume various roles, such as the role of a
student, teacher, mentor, or peer. Its successful implementation relies on carefully
crafted prompts that are clear, specific, contextualized, and adaptable. Effective strat-
egies include “chain-of-thought” prompts to encourage step-by-step problem-solv-
ing, “negative prompts” to prevent errors, and “meta-prompts” that guide students
toward reflection and critical thinking. Although this article focuses on applications
in mathematics, the findings are equally applicable to other subjects and broader
educational contexts.
Keywords: generative artificial intelligence (GEN-AI), mathematics education, prompt
design, personalized learning, learning mathematics

245