An Examination of Mathematics Pre-Service Teachers’ Anxiety Toward Artificial Intelligence and Their Learning Intentions

Abstract

This study aims to examine pre-service teachers’ intentions and anxiety levels regarding learning about artificial intelligence according to different demographic characteristics and to reveal the relationship between these variables. The sample of the research consists of 205 mathematics pre-service teachers studying in the Department of Mathematics Education at a state university. Students completed two 5-point Likert-type scales, one with 15 items and the other with 16 items, as well as a form containing open-ended questions. The study employed a sequential mixed methods design. As a result of the study, pre-service teachers had a high level of intention to learn about artificial intelligence (AI) and a moderate level of anxiety regarding AI was determined. Additionally, a weak but statistically significant negative correlation was found between learning intention and anxiety. In addition to this, it was concluded that pre-service teachers see artificial intelligence integration as an inevitable professional necessity in their future professional lives; that is, pre-service teachers’ concerns about artificial intelligence do not constitute a major barrier to their learning intentions.

1. Introduction

Rapid technological advancements are leading to significant changes in many fields. Among these advancements, artificial intelligence (AI) technologies have become increasingly integrated into various aspects of human life. However, it is stated that artificial intelligence continues to evolve by examining human intelligence [1]. Coppin [2] defines artificial intelligence as the ability of machines to adapt to situations they encounter, analyze problems and generate solutions, answer questions, and execute other functions requiring a certain level of intelligence. On the other hand, Nabiyev [3] considers artificial intelligence a system that develops solutions to complex problems and incorporates reasoning, decision-making, and learning processes. With the widespread use of artificial intelligence in various fields in recent years, these technologies have also become increasingly prevalent in education and training processes.

The rapid increase in the integration of artificial intelligence in education has significantly impacted education and training processes [4]. Artificial intelligence has been shown to transform many classroom activities, from teachers’ lesson planning and the instructional processes to student assessment and evaluation [5]. Instead of eliminating the role of teachers, artificial intelligence is considered a tool that contributes to teachers taking on more strategic and guidance-oriented responsibilities in pedagogical processes [6]. Through artificial intelligence, teachers can analyze students’ learning needs more accurately and adapt their instructional processes accordingly based on the data obtained. In addition, student development can be regularly monitored, academic achievement can be evaluated, and individual support can be offered to students when needed through artificial intelligence-based tools [7]. Teachers who actively use artificial intelligence in their lessons are seen to prefer it extensively for tasks such as question preparation, content creation, activity design, data analysis, and student achievement tracking [7]. Similarly, in another study [8], it was determined that artificial intelligence is not a tool that replaces the teacher, but rather a tool that supports teachers’ skills and contributes to more effective learning processes. However, it is stated in the literature that artificial intelligence also introduces several pedagogical and ethical concerns [9]. When the main concerns of teachers regarding this process are examined, issues such as the possibility that the use of artificial intelligence may negatively affect students’ creativity [10], the possibility that access to readily available information may lead students to cognitive laziness [11], inequalities in access to technology, and high costs [7] come to the forefront. These concerns may contribute to the development of negative attitudes toward artificial intelligence. It has also been shown that anxiety levels regarding artificial intelligence can affect pre-service teachers’ attitudes towards these technologies [12,13].

Therefore, it is important to promote more positive attitudes toward artificial intelligence and address negative perceptions. Furthermore, it is stated that increasing pre-service teachers’ knowledge and skills regarding artificial intelligence technologies is critically important for the success of digital transformation in education [14]. The positive aspects of artificial intelligence in education are clearly visible, but it is also stated that the effective use of these technologies in educational environments is largely related to teachers’ motivation to learn about and use learning artificial intelligence [15]. Positive motivation may contribute to the development of positive attitudes, and Erol and Erol (2024) [16] state in their study that teachers’ positive attitudes towards artificial intelligence technologies support the effective use of these technologies in educational environments. Accordingly, the study suggests that improving teachers’ attitudes towards artificial intelligence in a positive direction may enhance the effectiveness of artificial intelligence applications in education. Özkul [17] aimed to determine the general attitudes of pre-service teachers towards artificial intelligence and their attitudes towards the use of artificial intelligence in education. According to the research results, pre-service teachers’ attitudes towards artificial intelligence were found to be at a moderate level; however, no statistically significant difference was found between attitudes based on department and gender variables.

In their study with pre-service teachers, Erdoğdu and Çakır [10] stated that some pre-service teachers expressed definitively negative views towards artificial intelligence, believing that this technology often does not always provide reliable information. In Özkul’s [17] study examining pre-service teachers’ attitudes towards artificial intelligence, it was observed that pre-service teachers highlighted potential risks associated with artificial intelligence, such as technological dependence, weakening of professional roles, and declines in cognitive processes. These findings may reveal that future teachers have developed a cautious approach not only toward the potential disadvantages of artificial intelligence but also toward the challenges that may arise from its use, despite the opportunities it offers.

Anxiety is defined as an emotional state characterized by negative feelings toward anticipated future events, arising from an individual’s experience of inner unease [18]. In this context, similar concerns have become increasingly prominent in discussions and evaluations of artificial intelligence technologies in recent years. Concerns are expressed about the development of artificial intelligence, and there are discussions that this technology may become difficult to control and have negative effects on individuals and society [19]. These discussions about artificial intelligence in society also affect pre-service teachers. Duan et al. [20] highlights that the rapid advancements in artificial intelligence are leading to widespread anxiety among college students about their future careers. In particular, the fact that pre-service teachers, who will be the teachers of the future, will be directly interacting with artificial intelligence technologies makes it important to examine their levels of anxiety regarding these technologies. In a study, Sarıkaya and Şakiroğlu [13] examined the anxieties of Turkish language pre-service teachers towards artificial intelligence in terms of various variables. According to the study findings, it was determined that the anxiety levels of Turkish-language pre-service teachers towards artificial intelligence were at a moderate level.

Akgün et al. [12] examined the demographic variables affecting the attitudes, literacy, and anxiety levels of science and mathematics pre-service teachers towards artificial intelligence. The findings revealed a moderately positive relationship between attitude scores and AI literacy levels, as well as a weak positive relationship between attitude scores and anxiety levels. In contrast, the relationship between literacy and anxiety was negative, but not statistically significant. It was also concluded that female students had higher anxiety levels. Yeniçeri and Kenan [21] aimed to determine the attitudes and anxiety levels of pre-service teachers toward artificial intelligence and to examine the factors affecting these levels. The study was conducted with pre-service teachers from different teacher training fields. The results show that male pre-service teachers had more positive attitudes toward artificial intelligence than female pre-service teachers, whereas female pre-service teachers had higher levels of anxiety. Furthermore, it was determined that pre-service teachers who had previous experience with artificial intelligence exhibited more positive attitudes toward this technology and lower levels of anxiety. These findings suggest that positive attitudes toward artificial intelligence and previous experience with the artificial intelligence may encourage an individual’s future efforts, which gives rise to the concept of learning intention. In this context, positive attitudes toward artificial intelligence may increase individuals’ interest in technology and, consequently, their intention to learn about it.

Artificial intelligence learning intention is considered a multidimensional construct encompassing individuals’ desire to acquire knowledge about, access, engage with, and learn from AI tools [22,23]. It is regarded as a crucial factor in fully realizing the potential of AI in education and improving students’ learning experiences, as teachers need to possess the necessary knowledge and skills to effectively utilize these technologies [24]. Scherer and Siddiq [25] state that students may have different motivations for learning AI technologies. Some may have intrinsic motivation due to their interest in AI, while others may have extrinsic motivations stemming from external factors, such as environmental pressure. Esiyok et al. [26] emphasize that intrinsic motivation may encourage students to spend more time engaging with AI. Li et al. [27] also conducted a study to develop a survey measuring university students’ motivation to learn AI. Measuring AI learning intentions is thought to contribute to promoting the effective use of technology in education. Therefore, examining pre-service teachers’ learning intentions toward AI is considered important. A review of the literature reveals that pre-service teachers’ intentions to learn about AI may play a significant role in the effective integration and use of these technologies in educational processes [17,28].

However, most existing studies address AI learning intention and AI anxiety separately, and research examining the relationship between pre-service teachers’ AI learning intentions and anxiety levels is limited. Considering the benefits and drawbacks of artificial intelligence applications used in education for students and teachers, pre-service teachers’ approaches to and perspectives on artificial intelligence are of great importance, as their attitudes toward these technologies may influence their willingness to adopt and use them in educational environments. In this context, it is considered important to examine pre-service teachers’ intentions to learn about artificial intelligence and their anxiety regarding these technologies simultaneously. It is anticipated that examining the relationship between these two variables will contribute to understanding pre-service teachers’ approaches to artificial intelligence. Accordingly, this research aims to examine the relationship between mathematics pre-service teachers’ intentions to learn about artificial intelligence and their anxiety levels. It is expected that the findings of this study will contribute to addressing an important gap in the literature.

Accordingly, the research question was determined as follows: “Is there a significant relationship between mathematics pre-service teachers’ intentions to learn about artificial intelligence and their anxieties about artificial intelligence?” In this context, the sub-problems of the study are

1. What is the level of mathematics pre-service teachers’ intentions to learn about artificial intelligence?

2. Do the mathematics pre-service teachers’ intentions to learn about artificial intelligence show significant differences according to gender, class level, and the program they are studying?

3. What is the level of mathematics pre-service teachers’ anxieties about artificial intelligence?

4. Do the mathematics pre-service teachers’ anxiety levels about artificial intelligence show significant differences according to gender, class level, and the program they are studying?

5. Is there a statistically significant relationship between the mathematics pre-service teachers’ intentions to learn about artificial intelligence and their anxiety levels?

6. What are the mathematics pre-service teachers’ views and anxieties about the use of artificial intelligence?

6.1. What are the mathematics pre-service teachers’ views on the contributions of the use of artificial intelligence tools?

6.2. What are the mathematics pre-service teachers’ views on their anxieties about using and learning artificial intelligence?

6.3. What are the mathematics pre-service teachers’ views on the effects of artificial intelligence on the future of the teaching profession?

2. Materials and Methods

2.1. Research Model

This research is a mixed-methods study aiming to examine the relationship between mathematics pre-service teachers’ intentions to learn artificial intelligence (AI) and their anxiety levels, as well as their opinions and concerns regarding the use of AI. In mixed-methods research, the integration of data enables a more robust understanding of the problem [29]. In this study, an explanatory sequential design, one of the mixed-methods research methods, was used. The explanatory sequential design is a mixed-methods design where quantitative data are collected and analyzed in the first stage of the research, and then a qualitative stage is used in the second stage to explain these quantitative results in more depth [30].

At this point, quantitative data were collected using scales to determine the AI learning intentions and AI anxiety levels of pre-service teachers. Based on the scores obtained from 15 pre-service teachers were selected for the qualitative phase, with five participants each representing low, medium, and high levels of AI learning intention and AI anxiety. The selection process was conducted to ensure diversity across different levels of AI learning intention and AI anxiety and to include participants representing different combinations of these characteristics. Thus, the qualitative interviews included participants with varying levels of AI learning intention and anxiety, allowing these variables to be examined from different perspectives and in greater depth. This approach was intended to provide a more comprehensive understanding of how AI anxiety and learning intentions are experienced at different levels. Qualitative data, which enabled an in-depth examination of the pre-service teachers’ views and concerns regarding the use of artificial intelligence, were collected through an interview form.

2.2. Research Population and Sample

This research was conducted with 205 first, second, third and fourth-year pre-service teachers studying in the Mathematics Education Department of a State University during the Fall semester of the 2025–2026 academic year, who voluntarily participated in the study. Among the 205 pre-service teachers, from whom quantitative data was collected, 15 participants were selected to collect qualitative data. The sample for the quantitative research was created using the simple random sampling method. Simple random sampling refers to the random selection of a sufficiently large sample from a population with specific characteristics that can represent the population [31]. The sample of this research was determined using the convenient sampling method, one of the purposeful sampling methods used in qualitative research. The convenient sampling method is defined as the researcher focusing on the easiest elements to reach from the target population to create the sample [32]. However, in order to consider the confidentiality of identity information, pre-service teachers were asked to answer the questions without writing their names.

2.3. Data Collection Tools and Data Collection

In this study, a demographic information form prepared by the researchers was used as the data collection tool to determine the demographic characteristics of pre-service teachers. In addition to this qualitative interview form the scales used are described below. The “Artificial Intelligence Learning Intention Scale” developed by Kaplan and Kurdal and the “Artificial Intelligence Anxiety Scale” developed by Akkaya et al. can be accessed by referring to references [24] and [19] respectively.

2.3.1. Artificial Intelligence Learning Intention Scale

The “Artificial Intelligence Learning Intention Scale” developed by Kaplan and Kurdal [24] was used to determine the artificial intelligence learning intentions of mathematics pre-service teachers. The items of the scale are in the form of a 5-point Likert scale. Comprising four dimensions, the scale consists of a total of 15 items. Within the current sample, the reliability coefficients of the sub-dimensions of the scale are 0.75 for knowledge, 0.76 for access, 0.75 for interest, and 0.79 for learning, respectively. The overall reliability level of the scale was determined as Cronbach’s alpha coefficient of 0.79. The fact that the reliability coefficients for both the dimensions and the total scale are above 0.70 indicates that the scale has a high level of reliability.

2.3.2. Artificial Intelligence Anxiety Scale

To determine the levels of artificial intelligence anxiety among pre-service teachers, the “Artificial Intelligence Anxiety Scale” developed by Akkaya et al. [19] was used. This scale, consisting of four dimensions, comprises a total of 16 items and uses a 5-point Likert-type rating scale. The minimum possible score on the scale is 16, and the maximum is 80. A participant’s score is interpreted as an increase in AI anxiety. Within the current sample, the reliability coefficients for the sub-dimensions of the scale are 0.90 for learning, 0.82 for job changing, 0.78 for sociotechnical blindness, and 0.96 for artificial intelligence data. The overall reliability level of the scale was determined to be 0.90 using Cronbach’s alpha.

2.3.3. Qualitative Interview Form

The semi-structured interview form consisting of open-ended questions was developed to learn about pre-service teachers’ views on their intentions to learn about artificial intelligence and their concerns regarding artificial intelligence. The qualitative interview questions prepared for the study are provided in Appendix A at the end of the manuscript. For this stage, pre-service teachers were given 30 min to answer the demographic information form and scales and were asked to express their true thoughts.

2.4. Data Analysis

After completing the quantitative section of the study, which involved collecting data using the “Artificial Intelligence Learning Intention Scale” and the “Artificial Intelligence Anxiety Scale”, the obtained data were first transferred to Microsoft Excel and subjected to necessary pre-processing. Statistical analyses were then performed on the data using IBM SPSS Statistics, Version 25.0 (IBM Corp., Armonk, NY, USA).

First, normality analysis was conducted on the data groups. In addition to the Kolmogorov–Smirnov test, the Shapiro–Wilk test was used to test for normal distribution, and skewness and kurtosis were examined. The results are presented in Table 1.

Table 1. Normality Test Results.

	Kolmogorov–Smirnov	Shapiro–Wilk	Skewness Statistic	Std Error	Kurtosis Statistic	Std Error
AI Learning Intention	0.014	0.006	0.149	0.169	0.944	0.337
AI Anxiety	0.083	0.232	0.116	0.169	0.557	0.337

It is stated that when the results of the Kolmogorov–Smirnov and Shapiro–Wilk tests are p < 0.05, they do not show a normal distribution and non-parametric tests should be used [33]. Examining Table 1, it is seen that according to the Kolmogorov–Smirnov and Shapiro–Wilk tests, the Kolmogorov–Smirnov result (p < 0.05, p = 0.014) and the Shapiro–Wilk result (p < 0.05, p = 0.006) of the artificial intelligence learning intention scale do not show a normal distribution, but the Kolmogorov–Smirnov result (p > 0.05, p = 0.083) and the Shapiro–Wilk result (p > 0.05, p = 0.232) of the artificial intelligence anxiety scale do show a normal distribution.

Based on these results, it was decided to examine the skewness and kurtosis values. Skewness and kurtosis values are calculated by dividing the skewness by its standard error, and results between −1.96 and +1.96 indicate a normal distribution [33]. For the artificial intelligence learning intention scale, the ratio of the skewness to the standard error was found to be 0.881, and the ratio of the kurtosis to the standard error was 2.80. For the artificial intelligence anxiety scale, the ratios of the skewness and kurtosis to the standard errors were calculated as 0.686 and 1.650, respectively.

When all these analyses are examined, it is seen that the results of the AI anxiety scale show a normal distribution, but the results of the AI learning intention scale do not show a normal distribution. Therefore, parametric tests were used for the AI anxiety scale, and non-parametric tests were used for the AI learning intention scale. In this context, the artificial intelligence learning intention levels of pre-service teachers were examined with the Mann–Whitney U test, a non-parametric test, according to the gender and program variables of the pre-service teachers, while the class level variable of the pre-service teachers was analyzed with the Kruskal–Wallis H test. The artificial intelligence anxiety levels of pre-service teachers were examined using an independent sample t-test for gender and program variables, whereas the class level variable was analyzed using the Welch test. The Spearman correlation coefficient was used to examine the relationship between artificial intelligence learning intention and artificial intelligence anxiety.

In the qualitative part of the study, which examined pre-service teachers’ opinions and anxieties regarding their AI learning intentions, the answers obtained from open-ended questions were analyzed using content analysis, a qualitative analysis method. In the content analysis process, the collected data are first conceptualized and organized within a logical framework, and then the main themes explaining the data are identified using these concepts [31]. During the data analysis, coding was done separately by each researcher, and then cross-checks were performed to reach a common understanding of the themes.

3. Results

This section presents the findings related to each sub-problem, based on data obtained from mathematics pre-service teachers, under separate headings.

3.1. Results Related to the First Sub-Problem of the Research

The first sub-problem of the research was stated as “What is the level of pre-service teachers’ intentions to learn about artificial intelligence?” The data regarding the overall average scores obtained from the Artificial Intelligence Learning Intention scale, which was applied to reveal the levels of pre-service teachers’ intentions to learn about artificial intelligence, are given in Table 2.

Table 2. Descriptive Statistical Findings Regarding Pre-service teachers’ Intention to Learn Artificial Intelligence.

	N	$\overline{\mathbf{X}}$	SD	Minimum	Maximum
Total	205	56.26	7.15	32	75

N: Number of observations, $\overline{\mathbf{X}}$: Mean, SD: Standard Deviation.

When Table 2 is examined, it is seen that the average score of pre-service teachers on the artificial intelligence learning intention scale is 56.26. The minimum score obtained from the research group was determined as 32, and the maximum score was determined as 75. In determining the artificial intelligence learning intention levels, the theoretical minimum (15) and maximum (75) scores of the scale were taken as the basis. Accordingly, it is seen that the average score obtained is in the high-level range according to the scale score ranges. Accordingly, it can be said that the learning intention levels of pre-service teachers towards artificial intelligence are at a high level.

3.2. Results Regarding the Second Sub-Problem of the Research

The demographic characteristics in the second sub-problem of the research, “Do pre-service teachers’ intentions to learn about artificial intelligence show significant differences according to gender, class level, and the program they are studying?”, were examined accordingly. The Mann–Whitney U test was applied to examine the gender variable, and the data obtained are presented in Table 3.

Table 3. Findings Regarding Pre-service Teachers’ Intentions to Learn Artificial Intelligence According to the Gender Variable.

Gender	N	$\overline{\mathbf{X}}$	Z	p
Female	162	101.92	0.507	0.612
Male	43	107.07

N: Number of observations, $\overline{\mathbf{X}}$: Mean, Z: Z statistic, p: p-value.

Table 3 shows that the Mann–Whitney U test results indicate no statistically significant difference in the total scores of pre-service teachers’ intention to learn artificial intelligence based on gender (Z = 0.612, p > 0.050). In other words, gender does not have a distinctive effect on pre-service teachers’ intention to learn artificial intelligence, and their levels of intention to learn artificial intelligence are similar across genders.

To test the second sub-problem question based on class level, the Kruskal–Wallis H test was applied, and the data obtained are presented in Table 4.

Table 4. Findings of Pre-service teachers’ Intention to Learn Artificial Intelligence According to the Class Level Variable.

Class Level	N	$\overline{\mathbf{X}}$	SD	χ2	p
1st grade	40	117.95	3	6.858	0.070
2nd grade	46	113.62
3rd grade	60	90.78
4th grade	59	98.58

N: Number of observations, $\overline{\mathbf{X}}$: Mean, SD: Standard Deviation, χ²: Chi-square, p: p-value.

Table 4 shows that a Kruskal–Wallis H test was conducted to determine whether the class level of a pre-service teacher has an effect on their intentions to learn artificial intelligence. The results show that there is no significant difference in the intentions of pre-service teachers to learn artificial intelligence in terms of class level (χ² = 6.858, p > 0.050).

Considering the program in which pre-service teachers are studying, the Mann–Whitney U test was applied to test the problem question. The data obtained from the Mann–Whitney U test are presented in Table 5.

Table 5. Findings Regarding Pre-service Teachers’ Artificial Intelligence Learning Intentions According to Program Variable.

Program	N	$\overline{\mathbf{X}}$	Z	p
Primary school mathematics teacher training program	122	99.90	−0.909	0.360
Secondary school mathematics teacher training program	83	107.56

N: Number of observations, $\overline{\mathbf{X}}$: Mean, Z: Z statistic, p: p-value.

Table 5 shows that the Mann–Whitney U test results indicate no statistically significant difference in the total AI learning intention scores between pre-service teachers studying in the secondary school mathematics teacher training program and those studying in the primary school mathematics teacher training program (Z = −0.909, p > 0.050). This finding reveals that the levels of intention to learn artificial intelligence among prospective teachers do not differ significantly based on the program in which they are enrolled and that there is a similar level of intention to learn artificial intelligence in both programs.

3.3. Results Regarding the Third Sub-Problem of the Study

The third sub-problem of the study, expressed as “ What is the level of pre-service teachers’ anxieties about artificial intelligence?”, aimed to determine the anxiety levels of pre-service teachers towards artificial intelligence. The data obtained from the analyses are presented in Table 6.

Table 6. Descriptive Statistics Findings Regarding the Artificial Intelligence Anxiety Levels of Pre-service teachers.

	N	$\overline{\mathbf{X}}$	SD	Minimum	Maximum
Total	205	49.30	11.03	16	80

N: Number of observations, $\overline{\mathbf{X}}$: Mean, SD: Standard Deviation.

According to Table 6, the total score of pre-service teachers regarding artificial intelligence anxiety was determined to be 49.30. It is observed that the minimum score obtained from the research group is 16 and the maximum score is 80. In determining the anxiety levels, the theoretical minimum (16) and maximum (80) scores of the scale were taken as the basis. Accordingly, it is seen that the average score obtained is in the medium level range according to the scale score ranges. In other words, it can be said that the anxiety levels of mathematics pre-service teachers towards artificial intelligence are at a medium level.

3.4. Results Regarding the Fourth Sub-Problem of the Study

The fourth sub-problem of the study was defined as “Do mathematics pre-service teachers’ anxiety levels about artificial intelligence show significant differences according to gender, class level, and the program they are studying?” The demographic characteristics were examined in order in the findings related to this problem.

To test the problem question regarding the artificial intelligence anxieties of pre-service teachers according to the gender variable, an independent sample t-test was applied. The Levene test results showed that the assumption of homogeneity of variances was met (p = 0.560); therefore, the results in the “Equal variances assumed” row are reported, and the obtained data are presented in Table 7.

Table 7. Findings Regarding the Artificial Intelligence Anxieties of Mathematics Pre-service Teachers According to the Gender Variable.

Gender	N	$\overline{\mathbf{X}}$	SD	df	Mean Difference	t	p	Cohen’s d
Female	162	50.47	10.7	203	5.452	2.928	0.004	0.503
Male	43	45.02	11.3

N: Number of observations, $\overline{\mathbf{X}}$: Mean, SD: standard deviation, df: degrees of freedom, t: t statistic, p: p-value.

According to Table 7, it was determined that the anxiety levels of pre-service teachers regarding artificial intelligence showed a statistically significant difference according to gender (t (203) = 2.928, p < 0.05).

In other words, gender affects the artificial intelligence anxiety levels of pre-service teachers, with female pre-service teachers reporting high levels. The average scores of female pre-service teachers on the artificial intelligence anxiety scale ($\overline{x}$ = 50.47) are higher than those of male pre-service teachers ($\overline{x}$ = 45.02). It is stated that an effect size can be defined as weak if the Cohen’s d value is less than 0.2, moderate if it is 0.5, and strong if it is greater than 0.8 [34]. Accordingly, the effect size of the significant difference between the artificial intelligence anxiety of pre-service teachers and their gender is moderate (Cohen’s d = 0.503). According to these results, it is seen that female pre-service teachers have higher levels of artificial intelligence anxiety than male pre-service teachers.

The Welch test was applied to test the fourth sub-problem question according to the grade level variable, and the data obtained are presented in Table 8.

Table 8. Findings Regarding Pre-service Teachers’ Concerns About Artificial Intelligence According to the Class Level Variable.

Welch	df1	df2	p
1.525	3	100.307	0.213

df1: degrees of freedom 1, df2: degrees of freedom 2, p: p-value.

Before the analysis, the assumption of homogeneity of variances was examined with the Levene test, and it was found that the variances were not homogeneous (L (3, 202) = 3.475, p = 0.017).

Since the assumption of homogeneity of variances was violated, Welch test results were conducted to examine difference between groups. The results revealed no statistically significant difference in anxiety scores between grade levels (W (3, 100.307) = 1.525, p = 0.213). Accordingly, the anxiety levels of pre-service teachers did not depend on their grade level.

Considering the program variable in which pre-service teachers are studying, an independent sample t-test was applied to test the problem question. The Levene test results showed that the assumption of homogeneity of variances was met (p = 0.690); therefore, the results in the “Equal variances assumed” row are reported, and the obtained data are presented in Table 9.

Table 9. Findings of Mathematics Pre-service teachers’ Artificial Intelligence Anxiety According to the Program Variable.

Program	N	$\overline{\mathbf{X}}$	SD	df	Mean Difference	t	p	Cohen d
Primary school mathematics teacher training program	122	51.07	10.6	203	4.302	−2.781	0.006	0.398
Secondary school mathematics teacher training program	83	46.77	11.1

N: Number of observations, $\overline{\mathbf{X}}$: Mean, SD: standard deviation, df: degrees of freedom, t: t statistic, p: p-value.

According to Table 9, it was determined that the anxiety levels of pre-service teachers regarding artificial intelligence differed statistically significantly depending on the program they were studying (t (203) = −2.781; p < 0.05). In other words, the artificial intelligence concerns of pre-service teachers differ according to the programs. This difference was observed to be more pronounced among pre-service teachers in the primary school mathematics teacher training program. The average scores of primary school mathematics teacher training programs on the artificial intelligence anxiety scale ($\overline{x}$ = 51.07) were higher than the average scores of mathematics pre-service teachers ($\overline{x}$ = 46.77). Accordingly, the effect size of the significant difference between the artificial intelligence anxiety of pre-service teachers and their programs is moderate (Cohen d = 0.398). Based on these results, it is seen that primary school mathematics teacher training programs have higher levels of artificial intelligence anxiety than secondary school mathematics teacher training programs.

3.5. Results Regarding the Fifth Sub-Problem of the Study

To analyze the fifth sub-problem of the study, “Is there a statistically significant relationship between pre-service teachers’ intentions to learn about artificial intelligence and their anxiety levels?”, the normality of the variables was examined. While the artificial intelligence anxiety variable showed a normal distribution, the artificial intelligence learning intention variable did not. Since at least one of the variables did not show a normal distribution, Spearman Correlation Coefficient analysis was used to examine the relationship between the variables, and the obtained data are presented in Table 10.

Table 10. Findings on the Relationship Between Artificial Intelligence Anxiety and Artificial Intelligence Learning Intention.

	N	rs	p
AI Anxiety–AI Learning Intention	205	−0.140	0.044

N: Number of observations, r_s: Spearman’s rank correlation coefficient, p: p-value.

Table 10 reveals a negative, very weak, and statistically significant correlation between pre-service teachers’ AI anxiety and their AI learning intentions (r_s = −0.140; p = 0.044). Accordingly, as participants’ AI anxiety levels increase, their intentions to learn AI tend to decrease; however, this correlation is quite weak.

3.6. Results Regarding the Sixth Sub-Problem of the Research

The sixth sub-problem, which constitutes the qualitative part of the research, is given as “What are the mathematics pre-service teachers’ views and anxieties about the use of artificial intelligence?” In this context, their views on the effects of artificial intelligence on the future of the teaching profession, their views on using and learning artificial intelligence, and their views on the contributions provided by the use of artificial intelligence tools were examined in 3 stages.

3.6.1. Views of Mathematics Pre-Service Teachers on the Contributions Artificial Intelligence Tools for Their Purposes of Use

The contributions provided by the use of artificial intelligence tools to mathematics pre-service teachers are grouped under three themes: “contributions to studying”, “contributions to practical situations”, and “no contribution”. The data obtained regarding the contributions of using artificial intelligence tools to pre-service teachers are presented in Table 11.

Table 11. Pre-service teachers’ Views on the Contributions Provided by Using Artificial Intelligence Tools.

Themes	Sub-Themes	Participants
Contributions to studying	In homework	PT1, PT6
	In solving problems	PT1, PT2, PT8, PT9, PT10, PT11, PT12
	In explaining topics I don’t understand	PT1, PT6, PT7, PT8, PT9, PT11
	In proofs	PT11
Contributions to practical situations	In summarizing	PT2, PT3
	For getting ideas	PT4, PT8, PT15
	In revisions	PT5
No contribution	I don’t use	PT13, PT14

Table 11 shows that the teachers most frequently mention the contribution of the tools to their studying. For example, the statement of a pre-service teacher coded PT6, “I use it at least 5–6 times a week to do homework or for topics I find difficult”, supports the contribution to their study habits. Similarly, the statements of PT9, “I ask them to explain the topic I don’t understand in a way I can understand, write down the example question and solve it; I use it very often when exams are approaching, otherwise I don’t use it much”, and PT8, “I don’t use it very often, but I use GEN AI tools for questions I can’t solve or for questions I can’t answer; I use it to ask questions that come to mind about topics I don’t know”, highlight the contribution they make to their studying by using the tools both for explaining topics they don’t understand and for solving problems. Furthermore, it is observed that pre-service teacher coded PT11 supports the idea that it contributes to studying by stating from different perspectives, with the statement “I use it in proofs, problem-solving, and understanding the subject”. The statements of pre-service teacher coded PT13, “I don’t use it”, and PT14, “I don’t use it because it harms my own creativity and thinking capacity”, support the theme that it does not contribute to their studying.

However, the statements of the mathematics pre-service teacher coded PT4, “I don’t have much confidence in mathematics, but I get ideas about the beginning and how I can progress in the teaching process”, and the pre-service teacher coded PT3, “Sometimes I use it to summarize long pages”, highlight the practical contributions of these tools regarding idea generation and summarization. Similarly, PT2’s statement, “I usually use it during exam periods; I ask questions and request that it explain, summarize, and simplify them with short tactics”, points to the contribution it provides for practical situations.

3.6.2. Mathematics Pre-Service Teachers’ Views About Their Anxieties Using and Learning Artificial Intelligence

The views of mathematics pre-service teachers regarding the concerns regarding artificial intelligence learning and usage are grouped under three themes: “Concerns stemming from the individual”, “Concerns stemming from artificial intelligence”, and “No concern”. The data obtained regarding the concerns experienced by pre-service teachers during the use of artificial intelligence are presented in Table 12.

Table 12. Pre-service teachers’ Views on Concerns Created by Artificial Intelligence Learning and Use.

Themes	Sub-Themes	Participants
No anxiety	No anxiety	PT2, PT3, PT4, PT5, PT6, PT8, PT11, PT12, PT13, PT14
Anxiety stemming from the individual	I will do something wrong	PT15
	I will ask the wrong question	PT15
Anxiety stemming from artificial intelligence	Is it giving wrong information?	PT7, PT9
	Is it solving the question incorrectly?	PT1

Table 12 shows that the vast majority of mathematics pre-service teachers (PT2, PT3, PT4, PT5, PT6, PT8, PT11, PT12, PT13, PT14) stated that they had no concerns about learning and using artificial intelligence. For example, the statements of pre-service teachers coded PT2, “I don’t experience any anxiety about learning, in fact, I find it enjoyable”, PT4, “I have no anxiety whatsoever”, and PT6, “I have no anxiety about this; most AIs are open source, so the necessary information can be obtained with a little research”, support the theme of having no anxiety. In addition to those who reported no anxiety, some pre-service teachers expressed concerns related to both their own experiences and the use of artificial intelligence. Pre-service teacher PT15, who expressed anxiety related to their own experience, emphasized their concern about both making mistakes and asking the wrong questions with the statement, “I have concerns about doing something wrong. I have concerns about not being able to ask the right question”. PT9, who expressed concerns related to artificial intelligence, indicated their anxiety about it providing incorrect information with the statement, “I worry about it teaching or giving incorrect information”. Furthermore, pre-service teacher PT1’s statement, “I worry about whether it will solve the questions incorrectly or whether it will give examples from a reliable source”, reflects concerns about the accuracy of AI-generated solutions and the reliability of the sources it provides.

3.6.3. Mathematics Pre-Service Teachers’ Views on the Effects of Artificial Intelligence on the Future of the Teaching Profession

The views of mathematics pre-service teachers regarding the impact of artificial intelligence on the future of the teaching profession are grouped under two themes: “Positive views” and “Negative views”. The data obtained regarding the opinions of pre-service teachers on the impact of artificial intelligence on the future of the teaching profession are presented in Table 13.

Table 13. Mathematics Pre-service teachers’ Views on the Impact of Artificial Intelligence on the Future of the Teaching Profession.

Themes	Sub-Themes	Participants
Positive Effects	It cannot replace the teaching profession.	PT1, PT4, PT9, PT10, PT13, PT14
	It can improve the teaching profession.	PT3, PT4, PT5, PT6
Negative Effects	It can negatively affect the teaching profession.	PT8, PT11,
	It can replace the teaching profession.	PT7, PT15

According to Table 13, the vast majority of mathematics pre-service teachers believe that artificial intelligence will positively influence the future of the profession, while maintaining that it cannot replace human teachers. For instance, pre-service teacher coded PT1 stated, “Frankly, I don’t think it will ever replace the teaching profession. Because even though it is helpful, it cannot think as deeply as the human brain”, and pre-service teacher coded PT10 stated, “I don’t think the teaching profession will disappear. A teacher is not just someone who transmits information, but someone who connects with students and influences their mistakes”. These statements support the idea that artificial intelligence cannot replace the teaching profession. They specifically stated that it cannot replace it because it cannot think deeply or make connections as deeply as the human brain. However, among the positive effects, pre-service teacher PT4’s statement, “I think the teaching profession will not disappear, but it can be improved with Gen AI. I find it quite useful, especially for individualized instruction”, supports the idea that it will improve the profession.

Among mathematics pre-service teachers who believe there are negative effects, the statement of the pre-service teacher coded PT7, “In the scenario where they start handling things we should be doing, it causes problems…”, indicates that the profession will be replaced. Similarly, the statement of PT11, “In general, I think that teaching could be negatively affected for many reasons, that more qualified teachers will be in demand, and that it will gradually lose its importance”, supports the idea that the profession will be replaced and negatively impacted.

4. Discussion

This study aimed to examine the relationship between mathematics pre-service teachers’ intentions to learn artificial intelligence (AI) and their anxiety levels. The study found that pre-service teachers had a high level of intention to learn about artificial intelligence (AI) and a moderate level of anxiety regarding AI. However, while AI learning intention did not vary significantly by grade level, gender, and program, AI anxiety showed a significant difference according to gender and program. Additionally, a weak but statistically significant negative correlation was found between learning intention and anxiety. Furthermore, pre-service teachers reported using AI for many different purposes, which may support the development of their practical skills and learning processes. The study also concluded that the majority of pre-service teachers did not experience anxiety about learning and using AI. Similarly, the majority expressed positive views and reported having no concerns about the future of the teaching profession. Overall, the findings suggest that mathematics pre-service teachers do not perceive AI technologies as a threat to the future of the teaching profession; rather, they view AI as a complementary tool that can enrich pedagogical processes and contribute to the transformation of the teacher’s role.

The study found that pre-service teachers had a high level of intention to learn artificial intelligence (AI). One reason for this high intention may be their awareness of the importance and necessity of AI, as well as the contexts and purposes for which it can be used. Indeed, it is noteworthy that pre-service teachers highlighted the contributions of AI use to their academic studies and practical situations. Considering their opinions that they use AI tools for completing assignments, understanding topics they find difficult, summarizing information, generating ideas, and solving mathematical problems and proofs, it can be suggested that they are generally aware of how and in which contexts AI can be used. However, only two pre-service teachers stated that the use of AI tools did not contribute and therefore they did not use them is an indicator of the high level of intention to learn AI. The fact that these participants reported that AI did not contribute to their learning and that they therefore did not use it may be associated with insufficient knowledge and limited experience with AI technologies. The intention to learn about the artificial intelligence scale is defined through the subdimensions of knowledge, access, interest, and learning [22]. For this reason, the literature indicates that pre-service teachers approach artificial intelligence with reservations due to their insufficient knowledge of it and inability to use these technologies effectively, and consequently, they adopt a traditional understanding of education [35].

In addition, the study examining pre-service teachers’ attitudes toward artificial intelligence found that their attitudes were at a moderate level [17]. Similarly, a study conducted with pre-service teachers revealed that the participants exhibited highly positive attitudes toward artificial intelligence [36]. Studies in the literature can be interpreted as an indicator that pre-service teachers have a positive attitude towards artificial intelligence in education and therefore intend to learn artificial intelligence. It is stated that the intention to learn artificial intelligence is influenced by concepts such as motivation and attitude [22,23]. Moreover, the moderate level of intention to learn artificial intelligence also suggests that pre-service teachers may not fully utilize AI-supported tools effectively. Lim [37] points out in his study that the effective use of AI-supported tools varies depending on their attitudes towards technology and their intention to learn AI.

In addition to this, it was determined that pre-service teachers’ intentions to learn about artificial intelligence did not show a statistically significant difference according to the gender, class level, and type of program they were studying. Furthermore, the finding that AI learning intentions did not show a significant difference based on the gender variable aligns with studies in the literature that also found no significant difference between the gender variable and attitudes toward AI [38]. Furthermore, Li et al. [27] demonstrated in their study that AI learning motivation mediates the positive relationship between supportive environments and participation in learning with AI. In the studies of Özyürek et al. [39], it was concluded that as teachers’ concerns about artificial intelligence increased, their positive attitudes decreased and negative attitudes increased. In other words, while this study points to the relationship between anxiety and attitude, it can be stated that a positive attitude toward AI will have a positive effect on the desire to learn AI. However, studies in the literature have found a significant relationship between attitude toward AI and gender [40,41] which contradicts the results of the current study. In their study with science and mathematics pre-service teachers, Akgün et al. [12]. report a significant difference in AI attitudes based on the department of study.

The study concluded that pre-service teachers have a moderate level of anxiety regarding artificial intelligence. Similarly, studies indicating that university students have moderate levels of anxiety about artificial intelligence support the findings of this research [42,43,44]. Concerns about artificial intelligence have also been reported to be at a moderate level among healthcare workers, indicating that such concerns are not limited to the field of education [45]. At the same time, it is noteworthy that the vast majority of pre-service teachers reported having no concerns about artificial intelligence, as indicated by the qualitative findings of the study. The reason for the moderate level of anxiety may be related to the fact that pre-service teachers believe that the rapid pace of development in artificial intelligence will increase and that artificial intelligence will replace the teaching profession in the future. Indeed, qualitative data indicate that the vast majority of pre-service teachers believe artificial intelligence will have a positive impact on the future of the teaching profession. In other words, they have little concern about the future of their profession. They emphasized that AI could contribute to the development of the teaching profession and would not replace teachers, suggesting that they have generally positive perceptions of the future of their profession. The reason for these positive thoughts may be that they are aware of the fact that teaching requires intense emotional labor rather than just transferring knowledge.

In the qualitative data, although the vast majority of pre-service teachers expressed positive views regarding the future impact of artificial intelligence on the teaching profession, some participants also drew attention to potential negative effects. In this context, it can be stated that pre-service teachers are somewhat anxious. Regarding negative opinions, some participants expressed concern that artificial intelligence could replace teachers and negatively affect the profession. Güneş [44] revealed in her study that pre-service teachers are concerned about artificial intelligence replacing the teaching profession. In contrast, research suggests that although artificial intelligence has achieved a certain degree of success in processes such as data analysis and instructional planning, it cannot replace teachers in affective areas such as empathy and value transmission [46]. Similarly, Felix [47] in his study showed that although artificial intelligence offers high efficiency in educational processes, it does not have the capacity to assume the human dimensions of teaching, such as empathy, values, and social awareness; it cannot have a physical presence in classrooms like a teacher and cannot provide ethical guidance. Furthermore, it is emphasized that AI would be more useful as a teacher’s assistant [48].

Similar situations are also discussed in the service industry. It is emphasized that AI can be quite useful for tasks that need to be done outside of education [49]. In contrast to the findings regarding pre-service teachers, A study has revealed that artificial intelligence anxiety is linked to growing concerns about technology-induced unemployment [50]. These results appear to be consistent with the findings of the current study. Qualitative data indicate that, in addition to concerns about the future of their profession, pre-service teachers reported other sources of anxiety related to both them and artificial intelligence. In fact, it is mostly observed that they refer to situations related to a person’s self-confidence or trust in artificial intelligence during its use. Similarly, in a study conducted with Religious Culture and Ethics teachers, it was determined that one of the concerns regarding artificial intelligence was the accuracy and reliability of the information it provides to individuals [44]. In this study, some pre-service teachers expressed concerns regarding the possibility of artificial intelligence providing incorrect information or generating erroneous solutions. These concerns may be related to the nature of mathematics, where accuracy and logical consistency are important. It can be said that pre-service teachers have difficulties trusting artificial intelligence. Even the slightest mistake shakes the trust of pre-service teachers in the AI tool, and this situation also triggers their anxieties. Furthermore, it is also noteworthy that pre-service teachers are anxious about making mistakes or asking inappropriate questions when interacting with AI. This finding indicates that some pre-service teachers may not feel sufficiently competent or confident when using artificial intelligence tools. This situation may stem from their limited knowledge and experience regarding artificial intelligence technologies. Similarly, when the opinions of pre-service teachers were examined, it was found that a large portion of the participants faced difficulties because of their limited knowledge of how to use AI tools [51]. This situation can create reluctance and reduce motivation to learn and use these technologies. Furthermore, it is emphasized that users’ difficulties in understanding AI tools may hinder their effective use of AI-based applications [52]. It is remarkable that the number of pre-service teachers who express that they do not have anxiety is high. This may be because pre-service teachers see AI as a part of daily life or they are comfortable not experiencing ethical situations that AI can create.

The study shows that pre-service teachers’ anxieties about artificial intelligence differ significantly based on gender and the program they are studying but not significantly based on class level. For gender, female pre-service teachers have more anxiety than male pre-service teachers. This difference may be due to female pre-service teachers wanting to approach rapidly developing artificial intelligence with more caution. A review of the literature reveals that Sarıkaya and Şakiroğlu [13] and Akgün [12], in their studies with pre-service teachers, found that anxiety about artificial intelligence differed significantly based on gender, with female pre-service teachers having more anxiety, which is consistent with the results of this study. Similarly, it was found that female pre-service teachers had higher anxiety levels compared to male pre-service teachers [21,42]. The reason why no significant difference emerged in pre-service teachers’ anxiety regarding artificial intelligence based on grade level can be explained by the fact that they are not fully aware of this field which is still somewhat new to all pre-service teachers and that artificial intelligence technologies have not yet been fully integrated into all levels of education. Similarly, studies that concluded that pre-service teachers’ anxieties about artificial intelligence did not differ significantly according to the class level variable support the findings of this research [13,32].

A weak, negative, and statistically significant correlation was found between pre-service teachers’ intentions to learn about artificial intelligence and their anxiety levels. This indicates that as pre-service teachers’ anxiety levels regarding artificial intelligence increase, their intentions to learn about artificial intelligence tend to decrease. The weak correlation may indicate that anxiety alone cannot fully explain the differences in prospective teachers’ intentions to acquire knowledge about artificial intelligence. An increase in a negative factor such as anxiety may be associated with lower self-confidence regarding learning artificial intelligence; consequently, this could reduce prospective teachers’ willingness to acquire knowledge about artificial intelligence. The literature indicates a significant negative correlation between university students’ attitudes and anxieties towards artificial intelligence [53]. Falebita [5] in a study with pre-service mathematics teachers, found a significant but weak correlation between pre-service teachers’ anxieties about artificial intelligence and their level of adoption of AI. Furthermore, it was stated that AI anxiety has a significant negative correlation with pre-service teachers’ intentions towards AI [54]. These studies are consistent with the results of the present study, as they also identified a significant negative relationship between pre-service teachers’ anxiety about artificial intelligence and their intention to learn about AI. However, it also shows that anxiety alone is not a determinant of learning intention. In other words, even if pre-service teachers experience anxiety, it does not eliminate their learning intention. The reason for this may be that other factors such as curiosity can also contribute to their learning intentions. This indicates that even if pre-service teachers have concerns about the future of their profession in different aspects, they are aware that they will have to use this technology when they start their careers. This parallels the result obtained from qualitative data, which shows that pre-service teachers are less anxious about the future of their profession. Nevertheless, the technological demands of the current era may encourage pre-service teachers to maintain a high intention to learn about AI despite experiencing some level of anxiety, particularly because of the potential benefits of AI for tasks such as classroom management and the preparation of teaching materials.

Indeed, the qualitative part of the study revealed that pre-service teachers actively use artificial intelligence for various purposes, including completing assignments, understanding topics they have difficulty with, summarizing information, generating ideas, reviewing content, and solving mathematical problems and proofs. In particular, pre-service teachers reported using AI for practical purposes, such as summarizing topics they do not understand and preparing study notes. These uses may reduce their workload and allow them to focus more directly on understanding and learning the content. In other words, AI may help pre-service teachers save time on routine or preparatory tasks and devote more attention to the comprehension process. Beyond using AI for studying from a student’s perspective, pre-service teachers also use it for various practical purposes related to their future professional roles. For example, using AI for lesson planning and idea generation may support the development of pre-service teachers’ professional thinking and instructional design skills.

In their study examining pre-service teachers’ perceptions of artificial intelligence Erdoğdu and Çakır [10] stated that pre-service teachers intended to use artificial intelligence for purposes such as preparing educational materials and creating presentations. Güneş [44] stated in her study that pre-service teachers expressed that artificial intelligence could be used in various areas of the educational process, such as activity design, lesson plan preparation, and material design. It can be stated that pre-service teachers emphasized that these uses contribute to their lesson preparation on the one hand and to practical applications on the other. Sumakul et al. [55] stated that teachers have a positive perception of the use of artificial intelligence in classrooms and that the integration of artificial intelligence into education can be effective for both teachers and students. Overall, the findings of the present study are consistent with previous research in the literature. Nevertheless, the study has a limitation in that its sample consisted of pre-service teachers enrolled in a specific program at a state university, which may limit the generalizability of the findings to pre-service teachers in other programs or institutions. Despite this limitation, pre-service teachers’ awareness of where and for what purposes AI can be used, as well as their ability to integrate it into their learning processes, is consistent with the high level of intention to learn about artificial intelligence identified in the quantitative findings. Furthermore, teachers’ intentions to learn about AI are considered a critical variable in terms of adopting AI-based teaching strategies, digitizing lesson content, and creating student-centered learning environments [7].

5. Recommendations

Based on the results of this research, the following recommendations can be made:

• To increase pre-service teachers’ intentions to learn about artificial intelligence, AI-related content can be integrated into teacher training programs.
• Informative activities can be conducted to reduce pre-service teachers’ anxiety levels regarding AI, focusing on the limitations, ethical dimensions, and role of these technologies in education.
• More detailed analyzes can be conducted by considering variables such as artificial intelligence learning intention, attitude, and artificial intelligence literacy.
• The research can be expanded to include pre-service teachers from different universities and different teacher training fields, increasing sample diversity and size.