Inglês (pdf)
Artigo em XML
Tradução automática
Enviar este artigo por email
Permalink
Introduction
Artificial intelligence technologies (chatbots) in modern day teaching and learning remains one of the latest challenges for policy makers in education (Mageira et al., 2022). The introduction of new AI chatbots to the teaching-learning process has the potential of transforming it to modern way of knowledge dissemination, communication and acquisition (Adiguzel et al., 2023).
The integration of chatbots and technologies into teaching and learning is the most groundbreaking innovations in the e-learning field which enabled the development and bring about efficient and innovative solutions to major teaching-learning problems (Fernoaga et al., 2018). Gonda and Chu (2019) asserted that chatbots like Google dialog flow chatbot can be integrated into offline and online content to solve the challenges in teaching activities. Liu et al. (2022) opined that AI chatbots improves students’ thinking ability, learning and knowledge acquisition skills and expectations in higher education.
Dimitriadis (2020) concluded that chatbots technology offers numerous services toward personalized and adaptive learning and by extension, it can also serves as virtual teaching assistants by relieving teacher of repetitive tasks. P et al. (2021) submitted that chatbots technology can address the diversified and logistics issues in teaching-learning process as faced by teachers and students in normal class which can also integrated into online platforms. Prananta et al. (2023) asserted that the use of ChatGPT in education offers opportunities in science learning, media and support and improves knowledge management.
Riza et al. (2023) advocated that inclusion of AI chatbots in e-learningprovide personalized service though with potential benefits and obstacles. Higher education students seems to be getting more comfortable with the usage of AI chatbots for learning due to their perceived convenience and enhanced performance (Malik et al., 2021). Every sector depends greatly on information and communication technology due to its efficient and effectiveness in service delivery with growing and acceptance of artificial intelligence in the field of ICT. The introduction of AI chatbots can enhance student learning experience in higher education, improve their productivity, aid their communication and also assist them in knowledge acquisition skills (Sandu & Gide, 2019).
Lin et al. (2023) highlighted the importance of chatbots to includes the provision of instant and automated responses and also improve students’ performance during the learning cycle. The future of science teaching and learning especially mathematics and mathematics-oriented subjects/courses will involve personalized learning experience, blended learning environment, data’s collection, organization and management literacy, computational thinking and statistics through the integration of artificial intelligence tools for thoughtful implementation and professional development (Supriyadi & Kuncoro, 2023).
A study conducted by Durak and Onan (2023) was centered on examination of research on the use of AI chatbots technology in the field of education. The authors considered 19 researched reports and papers that were related to AI chatbots technology in Google scholar. The analysis of study focused on article subject matter, purpose, research method, discussion and recommendation. The result of the study revealed that most of the studies were focused on the use of chatbots that were integrated into telegram, WhatsApp, Slack and Facebook. The finding further shows that most of analyzed articles were carried out at higher education institutions. West (2023) carried out a search on AI-ChatGPT chatbots versions 3.5 and 4.0 and understanding of force concept inventory in an introductory physics course. The results from concluded that AI-ChatGPT chatbot responded to force concept inventory questions exactly physicist might answer the question.
The present study assessed the In-service and Pre-service physics teachers’ attitude toward artificial intelligence (exogenous variable having three first order constructs), technology readiness index (exogenous variable having four first order construct) and willingness to integrate AI physics enabled chatbots in the teaching and learning processes (endogenous variable) as shown in Figure 1.
Literature review and development of research hypotheses
The studies reviewed in this study were on previous empirical and systematic reviews’ reports and findings on attitude towards artificial intelligence and technology readiness index.
Attitude towards artificial intelligence
Suh and Ahn (2022) developed and validate a model and its measuring scales of attitude toward artificial intelligence. The scholars believes that attitude of student toward artificial intelligence determine the adoption and willingness to integrate it into teaching and learning processes. This model as developed and validated by the scholars were carried out using confirmatory and exploratory factor analysis and it comprises of three main constructs (Behavioural component, Cognitive component and Affective component) with items and scale that can be adopted and adapted to measure them.
Chiu et al. (2021) posited that perceptions of AI’s cognitive and operational capabilities positively relate to affective and cognitive attitudes. Schepman and Rodway (2020) asserted that general attitudes towards AI can be predicted through individual comfortableness with specific applications and latest technologies.
The component of human attitude on how to acts or behave towards an object, issue and situation is well described as a behavioral component of an attitudes. Wolf et al. (2020) asserted that attitude comprises of cognitive, affective and behavioural components that explains the cognitions, belief and emotional reactions, interest and interest of present and past actions. Conner et al. (2020) opined that inconsistency in cognitive and affective component of an attitude serves as overall determinant of behaviour.
HO1. Behavioural, Cognitive and Affective components of attitude towards AI significantly related to willingness to integrate AI physics enabled chatbots into teaching-learning process;
Technology readiness index
Blut and Wang (2019) asserted that technology readiness index as a two-dimensional construct (motivator and inhibitor) indirectly influence the use of technology through technology acceptance model and quality-value-satisfactions chain. Parasuraman and Colby (2015) opined that the recent streamlined and update in technology readiness index tagged technology readiness index (TRI 2.0) is a valid, reliable and can be useful in measuring people’s willingness to embrace and use cutting edge technologies like AI.
Parasuraman (2000) developed a model tagged technology readiness index with the scholar thought of its essential roles the model plays in marketing service. Individual’ s readiness to make use of new technologies can be measured by technology readiness index. This model is a two-dimensional construct which involves the motivator and inhibitor. The motivator construct of technology readiness index is divided into optimism and innovativeness. The optimism parts of technology readiness index work on the positive point of view of the respondent about the latest technology like AI tools and chatbots which make learning more effective. The innovativeness part of technology readiness index’s motivator explains the likelihood of an individual to adopt latest technologies. The second construct of the technology readiness index named inhibitor which is also divided into discomfort and insecurity. The insecurity construct of technology readiness index’ inhibitor explains the concern and risks latest technologies users been exposed to. These risks involve privacy infringement, security and general distrust of technology. The discomfort construct explains problems that may arise as a result of being overwhelmed by latest technologies.
Barkirtas and Akkas (2020) concluded that optimism construct of technology readiness index has a positive relationship on both consumer’s perceived usefulness and perceived ease of use while innovativeness construct has positive relationship with consumer’s perceived ease of use. Jarrar et al. (2020) concluded in a study sought to determine the technology readiness index’s effect on the adoption of InDubai application, that the motivator constructs (optimism and innovativeness) as modelled by Parasuraman can prove the individual intentions to adopt latest technologies while the inhibitor (insecurity and discomfort) can hinder the adoption of latest technologies.
In another study conducted by Julian and Dhini (2022), the result of the study revealed that optimism, innovativeness and discomfort components of technology readiness index significantly influenced the perceived ease of use and perceived usefulness of smartwatch use.
HO2. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to optimism (motivator) component of technology readiness index;
HO3. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to innovativeness (motivator) component of technology readiness index;
HO4. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to discomfort (inhibitor) component of technology readiness index;
HO5. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to insecurity (inhibitor) component of technology readiness index;
HO6. Optimism, Innovativeness, Discomfort and Insecurity components of technology readiness index significantly related to willingness to integrate AI Chatbots in teaching-learning process.
Research question
RQ1. Does difference exist in relationship between in and pre-service physics teachers’ attitude towards AI and willingness to integrate AI physics enabled chatbots in teaching-learning process?
RQ2. Which of the attitude towards AI and Technology readiness index mostly predict the respondents’ willingness to integrate AI physics enabled chatbots in teaching-learning process?
Method and measurement
The study uses the primary data gathered through face-to-face questionnaire administration to the respondents (pre-service and in-service physics teachers) that cut across the teachers’ training college and teachers’ training faculty at university and physics teachers who graduated, employed and engaged to teach at secondary school level in the last five years. The choice of the respondents was based on the fact that the popularity of artificial intelligence technologies/ tools/chatbots usage among the students is on the high sides in the last few years. According to Wikipedia, the investment and interest in AI gain momentum and popularity when machine learning tools were launched and adopted to solve many problems associated with academics, health, engineering, business and security. Modern correlation method was adopted to determine the complex relationship among attitude towards artificial intelligence model, technology readiness index model and willingness to integrate AI chatbots into teaching and learning process. In determine the variables relationship, influence and impact, the existing model of technology readiness index by Parasuraman (2020) with four constructs (optimism, innovativeness, discomfort and insecurity), attitude towards artificial intelligence model by Suh and Ahn (2022) with three constructs (Behavioural, Cognitive and Affective) were adopted and willingness to integrate AI chatbots into teaching-learning process. The constructs’ items were described in the Table 1. The variables involved in this study is complex and the modern relationship method known as structural equation model and machine learning were adopted in this study. The data collected were analyzed based on variables’ relationship and importance performance. SmartPLS version 4.0.9.2 software was used to determine the relationship among variables and Artificial Neural Network (ANN) algorithm of machine learning embedded in SPSS software was used to determine the importance performance of the exogenous variables to willingness to integrate AI physics enabled Chatbots in teaching-learning process.
Table1. Construsts’ Description.
| Second Order Latent Variable | First/Higher Order Latent Variable | Construct’s Meaning | Items’ Code | Construct’s Items | Validity Index | Source |
| Attitude towards AI | Thinking or feeling about AI | BE1 | It is fun to learn about AI | |||
| Behavioural Component | BE2 | It is interesting to use AI | >0.75 | Items adapted from Suh and Ahn (2022) | ||
| BE3 | I want to make something that makes human life more convenient | |||||
| BE4 | I think that there should be more class time devoted to AI in school | |||||
| Cognitive Component | COG1 | I think it is important content to learn about AI in school | ||||
| COG2 | I think that AI should be taught in school | >0.75 | Items adapted from Suh and Ahn (2022) | |||
| COG3 | I think every student should learn about AI in school | |||||
| COG4 | AI class is important | |||||
| Affective component | AFF1 | AI is related to my life | ||||
| AFF2 | I will use AI to solve problems in daily life | >0.75 | Items adapted from Suh and Ahn (2022) | |||
| AFF3 | AI is worth studying | |||||
| AFF4 | AI is very important for developing society | |||||
| Technology Readiness Index | Readiness to exploit opportunities’ offered by latest technology | OPT1 | Technology gives me more freedom of mobility | >0.75 | Items adapted from Parasuraman (2000) | |
| Optimism (Motivator) | OPT2 | Technology gives people more control over their daily life | ||||
| OPT3 | Technology makes me more efficient in my occupation and study | |||||
| OPT4 | I feel confident that technology-based systems will follow through with what I instruct them to do | |||||
| INN1 | Learning about technology can be as rewarding as the technology itself | |||||
| Innovativeness (Motivator) | INN2 | I enjoy the challenge of figuring out high-tech gadgets | >0.75 | Items adapted from Parasuraman (2000) | ||
| INN3 | I keep up with the most available advanced technology | |||||
| INN4 | Other people come to me for advice on new technologies | |||||
| DIS1 | I have fewer problems than other people in making technology work for me | |||||
| Discomfort (Inhibitor) | DIS2 | Sometimes, I think that technology systems are not designed for use by ordinary people | >0.75 | Items adapted from Parasuraman (2000) | ||
| DIS3 | Technology always seems to fail at the worst possible time | |||||
| DIS4 | Many technologies have health or safety risks | |||||
| Insecurity (Inhibitor) | INS1 | I worry that information I make available over internet maybe misused by others | >0.75 | Items adapted from Parasuraman (2000) | ||
| INS2 | I do not consider it safe to provide personal information over the internet | |||||
| INS3 | New technologies make it too easy to spy | |||||
| Willingness to integrate AI physics chatbots to teaching-learning process | The state of being prepared and ready to use AI chatbots for teaching and learning | WILL1 | I am willing to use physics Chatbots in classroom | |||
| WILL2 | I will recommend that others should use AI physics Chatbots in classroom | >0.75 | Items adapted from Chatterjee and Bhattacharjee (2020) | |||
| WILL3 | I am willing to use AI technology for developing physics content for teaching-learning process |
Demographic profile of the respondents
The Table 2 described the demographic profiles of the respondents. 100 respondents were engaged to participate in this study and were selected through non-probability sampling technology (Purposive Sampling). The choice of the respondents was necessitated by their access to smart phone, ICT gadgets and internet network. 19 respondents representing 19% were female while 81 respondents representing 81% were male. The categories of the respondents involve 45 in-service physics teachers representing 45.0% of the total respondents’ size and 55 pre-service physics teachers representing 55.0%.
Results and findings
Measurement model
In this context, the values in matrix format represent the HTMT ratio which is used to evaluate the extent to which the each construct discriminates from other constructs in the formed model as shown Tables 3, 4, and 5.
Heterotrait-monotrait (HTMT) of correlations of the constructs’ (discriminant validity)
Table 3 Complete table for both In and Pre-service physics teachers’ responses on constructs in the model.
| Complete | ||||||||
| Construct | A | B | C | D | E | F | G | H |
| A | ||||||||
| B | 0.516 | |||||||
| C | 0.774 | 0.823 | ||||||
| D | 0.371 | 0.155 | 0.224 | |||||
| E | 0.605 | 0.225 | 0.389 | 0.682 | ||||
| F | 0.094 | 0.035 | 0.083 | 0.769 | 0.232 | |||
| G | 0.816 | 0.457 | 0.635 | 0.469 | 0.771 | 0.139 | ||
| H | 0.172 | 0.353 | 0.106 | 0.128 | 0.172 | 0.226 | 0.178 |
Table 4 Complete table for both In -service physics teachers’ responses on constructs in the model.
| In-service | ||||||||
| Construct | A | B | C | D | E | F | G | H |
| A | ||||||||
| B | 0.543 | |||||||
| C | 0.778 | 0.885 | ||||||
| D | 0.306 | 0.165 | 0.169 | |||||
| E | 0.589 | 0.201 | 0.350 | 0.609 | ||||
| F | 0.059 | 0.067 | 0.022 | 0.760 | 0.134 | |||
| G | 0.837 | 0.469 | 0.632 | 0.423 | 0.748 | 0.049 | ||
| H | 0.180 | 0.351 | 0.112 | 0.162 | 0.177 | 0.229 | 0.130 |
Table 5 Complete table for both Pre-service physics teachers’ responses on constructs in the model.
| Pre-service | ||||||||
| Construct | A | B | C | D | E | F | G | H |
| A | ||||||||
| B | 0.503 | |||||||
| C | 0.769 | 0.739 | ||||||
| D | 0.456 | 0.156 | 0.293 | |||||
| E | 0.625 | 0.273 | 0.445 | 0.778 | ||||
| F | 0.236 | 0.058 | 0.174 | 0.777 | 0.346 | |||
| G | 0.793 | 0.457 | 0.640 | 0.522 | 0.801 | 0.259 | ||
| H | 0.171 | 0.368 | 0.104 | 0.194 | 0.171 | 0.322 | 0.231 |
A-Affective component of the attitude toward AI, B-Behavioural Component of the attitude toward AI, C-Cognitive Component of the attitude toward AI, D-Discomfort component of Technology Readiness Index, E-Innovativeness component of Technology Readiness Index, F- Insecurity component of Technology Readiness Index, G-Optimism component of Technology Readiness Index and H-Willingness to integrate AI- chatbots in teaching-learning process.
Convergent validity
The Tables 6 and 7 below contains the various reliability and validity indexes of the measured constructs in the model. Cronbach Alpha values measures the internal consistency and by extension the extent to which items of a scale or constructed are correlated. The Cronbach Alpha’s value closer to 1 indicate stronger internal consistency. The composite reliability (rho_a and rho_c) are also alternative means of calculating the internal consistency of the constructs.
Average variance extracted (AVE) measures the amount of variance captured by the construt in relation to the amount of variance due to measurement error. A higher AVE value equal or above 0.5 indicated significant validity index. The Table5 shows the reliability and validity indexes of the combined In and pre-service responses on the constructs in the model while Table6 shows the separate reliability and validity indexes of In and Pre-service teachers’ responses since the study is on multi-group analysis.
Table 6 Complete table for both In and Pre-service physics teachers’ responses on constructs in the model.
| Complete | ||||
| Construct | Cronbach Alpha | Composite Reliability Rho_a | Composite Reliability Rho_c | Average Variance Extracted (AVE) |
| A | 0.920 | 0.933 | 0.943 | 0.805 |
| B | 0.605 | 0.728 | 0.783 | 0.515 |
| C | 0.889 | 0.905 | 0.922 | 0.749 |
| D | 0.862 | 0.897 | 0.914 | 0.781 |
| E | 0.955 | 0.959 | 0.967 | 0.881 |
| F | 0.915 | 0.938 | 0.946 | 0.854 |
| G | 0.956 | 0.961 | 0.968 | 0.883 |
| H | 0.834 | 0.873 | 0.897 | 0.743 |
Table 7.Tables for both In and Pre-service physics teachers’ responses on constructs in the model.
| In-service | ||||
| Construct | Cronbach Alpha | Composite Reliability Rho_a | Composite Reliability Rho_c | Average Variance Extracted (AVE) |
| A | 0.925 | 0.935 | 0.947 | 0.816 |
| B | 0.637 | 0.777 | 0.793 | 0.535 |
| C | 0.896 | 0.926 | 0.926 | 0.759 |
| D | 0.853 | 0.920 | 0.901 | 0.754 |
| E | 0.965 | 0.968 | 0.974 | 0.904 |
| F | 0.892 | 0.960 | 0.919 | 0.792 |
| G | 0.952 | 0.958 | 0.965 | 0.875 |
| H | 0.791 | 0.798 | 0.877 | 0.705 |
| Pre-service | ||||
| Construct | Cronbach Alpha | Composite Reliability Rho_a | Composite Reliability Rho_c | Average Variance Extracted (AVE) |
| A | 0.913 | 0.931 | 0.939 | 0.793 |
| B | 0.555 | 0.673 | 0.764 | 0.484 |
| C | 0.881 | 0.889 | 0.918 | 0.737 |
| D | 0.870 | 0.880 | 0.921 | 0.795 |
| E | 0.942 | 0.948 | 0.958 | 0.852 |
| F | 0.931 | 0.936 | 0.956 | 0.879 |
| G | 0.961 | 0.965 | 0.972 | 0.896 |
| H | 0.880 | 0.931 | 0.922 | 0.798 |
A-Affective component of the attitude toward AI, B-Behavioural Component of the attitude toward AI, C-Cognitive Component of the attitude toward AI, D-Discomfort component of Technology Readiness Index, E-Innovativeness component of Technology Readiness Index, F- Insecurity component of Technology Readiness Index, G-Optimism component of Technology Readiness Index and H-Willingness to integrate AI- chatbots in teaching-learning process.
Structural model
Testing of the Research Hypotheses
HO1: Behavioural, Cognitive and Affective components of attitude towards AI significantly related to willingness to integrate AI physics enabled Chatbots into teaching-learning process;
The interpretation of the Table 8 above indicated that the relationship between the three components of attitude towards AI and willingness to integrate AI chatbots into teaching-learning process are low, moderate, positive (Affective & Behavioural) and negative (Cognitive) though all the components are significantly related.
HO2. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to optimism (motivator) component of technology readiness index;
Table 8 Complete (In-service and Pre-service) Coefficient Table of Attitude towards AI and Willingness to integrate AI chatbots in teaching-learning process.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> WILL | 0.211 | 0.209 | Positive/Low | 2.438 | 0.015 | Supported |
| BE -> WILL | 0.367 | 0.364 | Positive/Moderate | 4.910 | 0.000 | Supported |
| COG -> WILL | -0.348 | -0.343 | Negative/Moderate | 4.139 | 0.000 | Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and WILL- Willingness to integrate AI chatbots in teaching-learning process.
The interpretation of the Table 9 above indicated that the relationship between the three components of attitude towards AI and optimism component of technology readiness index are low, substantial, positively related though the relationship between behavioural and cognitive components attitude towards AI with optimism component of technology readiness index were not significant.
HO3. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to innovativeness (motivator) component of technology readiness index;
Table 9 Complete (In-service and Pre-service) Coefficient Table of Attitude towards AI and Optimism component of technology readiness index.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> OPT | 0.717 | 0.715 | Positive/Substantial | 15.986 | 0.000 | Supported |
| BE -> OPT | 0.052 | 0.054 | Positive/Low | 1.298 | 0.194 | Not Supported |
| COG -> OPT | 0.057 | 0.057 | Positive/Low | 0.992 | 0.321 | Not Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and OPT- Optimism Component of Technology Readiness Index.
The interpretation of the Table 10 above indicated that the relationship between the three components of attitude towards AI and innovativeness component of technology readiness index are low, substantial, positively (Affective component) and negatively (Behavioural and Cognitive components) related though the relationship between behavioural and cognitive components attitude towards AI with innovativeness component of technology readiness index were not significant.
HO4. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to insecurity (inhibitor) component of technology readiness index;
Table 10 Complete (In-service and Pre-service) Coefficient Table of Attitude towards AI and Innovativeness component of technology readiness index.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> INN | 0.622 | 0.621 | Positive/Substantial | 11.681 | 0.000 | Supported |
| BE -> INN | -0.050 | -0.048 | Negative/Low | 0.950 | 0.342 | Not Supported |
| COG -> INN | -0.033 | -0.034 | Negative/Low | 0.507 | 0.612 | Not Supported |
AFF - Affective Component of the attitude toward AI, BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and INN- Innovativeness Component of Technology Readiness Index.
The interpretation of the Table 11 above indicated that the relationship between the three components of attitude towards AI and insecurity component of technology readiness index are low, positively (Affective & Cognitive components) and negatively (Behavioural component) related though the relationship between the constructs were not significant.
HO5. Behavioural, Cognitive and Affective component of attitude towards AI significantly related to discomfort (inhibitor) component of technology readiness index;
Table 11 Complete (In-service and Pre-service) Coefficient Table of Attitude towards AI and Insecurity component of technology readiness index.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> INS | 0.075 | 0.076 | Positive/Low | 1.034 | 0.301 | Not Supported |
| BE -> INS | -0.044 | -0.042 | Negative/Low | 0.606 | 0.545 | Not Supported |
| COG -> INS | 0.054 | 0.053 | Positive/Low | 0.616 | 0.538 | Not Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and INS- Insecurity Component of Technology Readiness Index.
The interpretation of the Table 12 above indicated that the relationship between the three components of attitude towards AI and discomfort component of technology readiness index are low, moderate, positively (Affective component) and negatively (Behavioural and Cognitive components) related though the relationship between behavioural and cognitive components of attitude towards AI with innovativeness component of technology readiness index were not significant.
HO6. Optimism, Innovativeness, Discomfort and Insecurity components of technology readiness index significantly related to willingness to integrate AI chatbots in teaching-learning process.
The interpretation of the Table 13 above indicated that the relationship between the four components of technology readiness index and willingness to integrate th AI chatbots into teaching-learning process are low, moderate, positively (Innovativeness & Insecurity component) and negatively (Optimism and Discomfort components) related though the relationship between optimism component of technology readiness index was not significantly related to willingness to integrate AI chatbots into teaching-learning process.
Table 12 Complete (In-service and Pre-service) Coefficient Table of Attitude towards AI and Discomfort component of technology readiness index.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> DIS | 0.408 | 0.410 | Positive/Moderate | 6.215 | 0.000 | Supported |
| BE -> DIS | -0.060 | -0.059 | Negative/Low | 0.831 | 0.406 | Not Supported |
| COG -> DIS | -0.042 | -0.043 | Negative /Low | 0.501 | 0.617 | Not Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and DIS- Discomfort Component of Technology Readiness Index.
Table 13 Complete (In-service and Pre-service) Coefficient Table of components of technology readiness index and willingness to integrate AI chatbots into teaching-learning process.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| OPT -> WILL | -0.027 | 0.027 | Negative/Low | 0.306 | 0.760 | Not Supported |
| INN -> WILL | 0.294 | 0.289 | Positive/Low | 2.980 | 0.003 | Supported |
| INS -> WILL | 0.457 | 0.460 | Positive /Moderate | 5.814 | 0.000 | Supported |
| DIS -> WILL | -0.432 | -0.431 | Negative /Moderate | 4.508 | 0.000 | Supported |
OPT- Optimism component of technology readiness index, INN- Innovativeness component of technology readiness index, INS- Insecurity component of technology readiness index, DIS- Discomfort component of technology readiness index and WILL- Willingness to integrate AI chatbots in teaching-learning process.
The interpretation of the Table 14 above shows the result of group A (In-service) and it indicated that the relationship between the three components of attitude towards AI and willingness to integrate AI chatbots into teaching-learning process are moderate, positive (Affective & Behavioural) and negative (Cognitive) though all the components are significantly related.
Table 14 In-service Coefficient Table of Attitude towards AI and Willingness to integrate AI chatbots in teaching-learning process.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> WILL | 0.315 | 0.307 | Positive/Moderate | 2.181 | 0.029 | Supported |
| BE -> WILL | 0.362 | 0.349 | Positive/Moderate | 2.900 | 0.004 | Supported |
| COG -> WILL | -0.359 | -0.339 | Negative/Moderate | 5.814 | 0.000 | Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and WILL- Willingness to integrate AI chatbots in teaching-learning process.
The interpretation of the Table 15 above shows the result of group B (Pre-service) and it indicated that the relationship between the three components of attitude towards AI and willingness to integrate AI chatbots into teaching-learning process are low, moderate, positive (Affective & Behavioural) and negative (Cognitive) though affective component of the attitude towards AI relationship with willingness to integrate AI chatbots into teaching-learning process was not significant.
Table 15 Pre-service Coefficient Table of Attitude towards AI and Willingness to integrate AI chatbots in teaching-learning process.
| Path | Path Coeff. (β) | Coeff. Mean | Remark | T-value | P-value | Remark |
| AFF -> WILL | 0.088 | 0.079 | Positive/Low | 0.784 | 0.433 | Not Supported |
| BE -> WILL | 0.382 | 0.380 | Positive/Moderate | 3.897 | 0.000 | Supported |
| COG -> WILL | -0.341 | -0.329 | Negative/Moderate | 3.128 | 0.002 | Supported |
AFF - Affective Component of the attitude toward AI,BE - Behavioural Component of the attitude toward AI,COG- Cognitive component of the attitude toward AI and WILL- Willingness to integrate AI chatbots in teaching-learning process.
Research questions
RQ1. Does difference exist in relationship between in and pre-service physics teachers’ attitude towards AI and willingness to integrate AI physics enabled chatbots in teaching-learning process?
Multi-group difference between In and Pre-service physics teacher’s willingness to integrate AI chatbots in teaching-learning process.
The differences are not significantly supported by both In and Pre-service physics teachers
The interpretation of the Table 16 above shows the result of multi-group difference (In-service - Pre-service) and it indicated that the relationship between the three components of attitude towards AI and willingness to integrate AI chatbots into teaching-learning process. The In-service physics teachers’ affective component of the attitude towards AI was higher than pre-service physics teachers though the other two components (Behavioural and Cognitive components) favours the Pre-service physics teachers and all the difference were not significant.
Table 16 In and Pre-service Path Coefficient of Attitude towards AI and Willingness to integrate AI chatbots in teaching-learning process difference.
| Path Relationship | Difference (In-service - Pre-service) | P-value | Remark |
| AFF -> WILL | 0.227 | 0.211 | Not Supported |
| BE -> WILL | -0.020 | 0.929 | Not Supported |
| COG -> WILL | -0.017 | 0.910 | Not Supported |
RQ2. Which of the attitude towards AI and Technology readiness index mostly predict the respondents’ willingness to integrate AI physics enabled Chatbots in teaching-learning process?
The research question two was answered using machine model called Neural Network. The variable importance Table 17 and Figure 2 below explains that the respondents’ attitude towards AI predicted and forecasted their willingness to integrate AI chatbots into teaching-learning process.
Table 17 Independent Variable Importance.
| Importance | Normalized Importance | |
| Attitude towards AI | .778 | 100.0% |
| Technology Readiness Index | .222 | 28.5% |
Figure 2 The Neural Network Graphical Output of the Importance of Independent Variables (Constructs).
The interpretation of the above Table 17 indicated that respondents’ attitude towards AI remain the most important factor that can aid the teachers to integrate the AI chatbots in their teaching-learning process.
Discussion (Separate topics)
This study explored the attitude toward artificial intelligence, technology readiness index and their relationship between In and Pre-service physics teachers’ willingness to integrate AI chatbots into teaching-learning process at secondary school physics class. The findings of this research show that the affective, behavioural and cognitive components of the attitude toward AI are significantly correlated with In and Pre-service physics teachers willingness to integrate AI chatbots in teaching-learning process. The first research hypothesis that was formulated and tested was to examine the correlation strength and significance of their correlation. The first hypothesis tested the relationship between attitude towards AI and In and Pre-service physics teachers’ willingness to integrate AI chatbots into teaching-learning process. The result in Table 8 revealed that affective component of the attitude towards AI had low and positive correlation coefficient with willingness to integrate AI chatbots, behavioural component had moderate and positive correlation with willingness to integrate AI chatbots while cognitive had moderate and negative correlation with willingness to integrate AI chatbots though all the three components of the attitude towards AI were all significantly correlated with willingness to integrate AI chatbots to teaching-learning process. The implication of this result means that the improvement in affective and behavioural components of the attitude towards AI will also improve the respondents’ willingness to integrate AI chatbots into teaching-learning process. The cognitive component of the attitude towards AI’ s relationship with respondents’ willingness to integrate AI chatbots into teaching-learning process revealed that the improvement in cognitive component of the attitude towards AI reduces the respondents’ willingness to integrate AI chatbots into teaching-learning process.
The second major formulated hypothesis which shows the linear relationship between technology readiness index’s components and respondents’ willingness to integrate AI chatbots into teaching-learning process. The results in Table8 revealed that optimism and discomfort components of technology readiness index had low, moderate, negative correlation with willingness to integrate AI chatbots into teaching-learning process though discomfort component significantly correlated with the willingness to integrate AI chatbots into teaching-learning process. The innovativeness and insecurity components of technology readiness index had low, moderate, positive correlation and significantly correlated with respondents’ willingness to integrate AI chatbots into teaching-learning process. The result explained that optimism (motivator) and discomfort (inhibitor) of the technology readiness index might not influence the respondents’ willingness to integrate AI chatbots teaching-learning process while the innovativeness (motivator) and insecurity (inhibitor) of the technology readiness index might influence the respondents’ willingness to integrate AI chatbots into teaching-learning process.
The first research question raised was centred on the multi-group analysis difference of the In and pre-service physics teachers’ attitude towards AI and their willingness to integrate AI chatbots into teaching-learning process. The results in Table15 revealed that In-service physics teachers’ affective component of the attitude towards AI was higher than pre-service physics teachers though the other two components (Behavioural and Cognitive components) favours the Pre-service physics teachers and all the differences were not significant.
The second research question raised was on the exogenous (independent) variables’ importance and their influence on the endogenous (willingness to integrate AI chatbots into teaching-learning process). The neural network result in Table16 explained that the respondents’ attitude towards AI remain the most important factor that might influence the decision of the respondents’ to willingly integrate and adopt the usage of AI chatbots in teaching-learning process.
Conclusion
The adoption and integration of the AI chatbots in classroom teaching-learning and communication process are fast growing and gaining momentum globally. The usage of AI chatbots by many professionals in enhancing their job output is on the rise and the educational sector is not left out. The result of the study concluded that attitude towards AI plays significant roles in adoption, integration, usage and recommendation of chatbots to solve human and machine related problems.
