Survey items for AI-readiness

1. Introduction

Artificial intelligence (AI) has been increasingly used in a variety of fields (e.g., industry, finance, and education) to promote innovation and increase work efficiency (Ng et al., 2021). In education, AI is touted as a seemingly almighty tool, supporting or even replacing teachers' work by automatically tracking students’ progress, assessing their performance, and providing personalized help (Albacete et al., 2019; Chounta et al., 2022; Tarus et al., 2018). Teachers can rely on AI to make informed decisions on orchestrating teaching practice so as to better support student learning (Van Leeuwen & Rummel, 2020).

Nonetheless, in reality, intelligent tools for education are rarely used consistently in K-12 classrooms (Ferguson et al., 2016). Schiff (2021) found that much practice and research related to the educational use of AI did not deliver promised changes and benefits. Among the multiple reasons leading to this controversy, for instance, the quality of AI and users' preferences (Luckin et al., 2022) and ethical concerns (Holmes et al., 2022), an essential culprit could be the techno-centric approach vehemently promoted by some in the educational field, which stresses the role of AI but ignores the agency of teachers who can decide whether, what, when, and how AI technologies are used in the first place (Luckin et al., 2022). Teachers are on the front lines of AI deployment, bridging schools' AI policies and students’ needs, thereby the critical role in the successful implementation of AI in schools (Felix, 2020). However, many teachers may not be actually ready for AI-enhanced education, though they are mostly aware of the potential benefits that AI can bring to education (Chounta et al., 2022). Their inadequate AI readiness may partially contribute to the gap between rapid advances in AI technologies and comparatively slow and unsatisfactory adoption of them in education (Luan et al., 2020; Luckin et al., 2022).

According to prior research on AI readiness (e.g., Holmström, 2022; Karaca et al., 2021; Luckin et al., 2022) while considering this study's context, AI readiness is defined as the state of preparedness among teachers in terms of their cognition, ability, vision, and ethical considerations with respect to the use of AI in education. Theoretically, teachers with high levels of AI readiness may have the knowledge and competence necessary for innovating their work by experimenting with and adapting to opportunities promised by AI (Jöhnk et al., 2021; Luckin et al., 2022). The innovative attempts may in turn improve their work experience, fostering high job satisfaction (Bhargava et al., 2021). Conversely, those with low levels of AI readiness may feel threatened, worrying about possible disruptions caused by AI to their work and subsequently alienating themselves from AI technologies (Chounta et al., 2022; Luckin et al., 2022).

Though a decent level of AI readiness is considered crucial for successful integration of AI into teaching (Celik et al., 2022), there is limited empirical knowledge regarding how AI readiness affects teachers' work. Even less is known regarding whether and how AI readiness may differ among teachers from distinct demographic backgrounds, particularly genders and socioeconomic backgrounds which have often been reported to cause disparities in the use of conventional technologies (Beaunoyer et al., 2020; Park et al., 2019). In addition, given that the ethical use of AI has been a concern attracting substantial attention (Hagendorff, 2020; Smakman et al., 2021), it would be useful to gain insights into how ethics are related to other components of AI readiness. Considering the increasing use of AI in education for innovating teaching and enhancing educators’ work experience (Celik et al., 2022; Luckin et al., 2022), this study sets out to bridge these gaps by addressing the following research questions:

RQ1

How are ethics related to other components of teachers' AI readiness?

RQ2

How are teachers' AI readiness associated with their perceived threats from AI, AI-enhanced innovation, and job satisfaction?

RQ3

How do teachers from different demographic backgrounds vary in AI readiness, perceived threats from AI, AI-enhanced innovation, and job satisfaction?

2. Literature review and hypotheses development

2.1. Prior research on AI readiness

The concept of AI readiness is relatively new, as highlighted in recent studies (Jöhnk et al., 2021; Luckin et al., 2022). Most of the research on AI readiness has been conducted in the business field, where AI has been adopted more widely than in education (Luckin et al., 2022). However, the components of AI readiness are still developing and may differ depending on the specific fields of application.

In the early days of technological readiness research, Parasuraman (2000) proposed a concept of technological readiness and defined it as individuals' inclination to adopt and use new technologies for completing tasks at home and work. Technology readiness was related to mental enablers and inhibitors that determined people's purchase decisions about new technologies. According to Parasuraman (2000), individuals' tendency to use new technologies was caused by the interplay between readiness enablers, including optimism toward technologies and innovativeness for their work, and readiness inhibitors, including discomfort and insecurity resulting from distrust of technologies. However, the technology readiness concept was developed for the service industry to improve customer satisfaction by identifying factors that decrease customers' frustration when interacting with new technologies and encourage their purchase intention. Compared with customers' readiness of using new technologies, educators' readiness of using AI is related to not only themselves but also their students. In addition, AI is different from conventional technologies in that it simulates some degree of human reasoning and learning while conventional technologies acquiesced full control to human beings (Damerji & Salimi, 2021). Therefore, the concept of AI readiness has to be reconsidered and redefined, particularly for educators.

Jöhnk et al. (2021) conceptualized organizational AI readiness based on interviews with 25 AI experts. It comprised 18 factors along five categories, including strategic alignment (aligning organizational needs with AI's potential), resources (finances, personnel, and IT infrastructure allocated for AI implementation), knowledge (AI awareness, skills, and ethics), culture (innovativeness, collaboration, and change management), and data (availability and quality of data for building valid AI models). Nonetheless, the AI readiness framework was developed for the industry sector and mainly reflected the views of the management staff. The 18 factors within five categories were obtained through qualitative analysis of experts' views and were not validated empirically through quantitative studies, thus leading to doubtful generalization to a broad population and other fields.

In education, Luckin et al. (2022) gave a more comprehensive introduction of AI readiness and highlighted contextualization when applying AI readiness to the educational sector. They proposed an AI readiness training framework comprising seven steps, involving engaging with the idea of AI readiness, pinpointing challenges in education to be solved, identifying and collecting data to address the challenges, applying AI techniques for data analysis, learning from the AI results, and iterating the framework if needed. However, Luckin et al.’s (2022) AI readiness framework was developed based on Cross-Industry Standard Process for Data Mining in the business sector, focusing on AI-supported data mining, instead of AI-enhanced teaching practice. In addition, applying the a priori framework from the business sector directly to the field of education may not well address educational challenges facing students and educators.

To prepare medical students for new roles and tasks in AI-enhanced healthcare, Karaca et al. (2021) developed an AI readiness scale for them, which comprised four components, including cognition (cognitive readiness regarding students' basic knowledge of AI), ability (students' competence in using AI for learning), vision (students' critical understanding of AI), and ethics (legal and ethical norms for responsible use of AI). However, Karaca et al. (2021)'s AI readiness scale was developed for specific healthcare student population and the empirical relationships between AI readiness and factors related to AI-enhanced learning were not reported. Despite these, among the prior studies examining AI readiness, Karaca et al. (2021)'s scale is particularly relevant to the research on AI readiness for educators, as their conceptualization of AI readiness is comprehensive and empirically validated.

In short, prior studies (e.g., Holmström, 2022; Luckin et al., 2022) have generally agreed on the importance of AI readiness for individual and organizational use of AI. Nevertheless, the studies on AI readiness are mostly conceptual, theorizing on its definition and factors comprising it. Though AI is gradually becoming a part of education (Brouillette, 2019), limited attention has been given to teachers who normally oversee the design and implementation of AI-enhanced education (Felix, 2020). Few studies have been conducted to empirically examine or validate the concept of AI readiness and its implications for teachers’ work (Bhargava et al., 2021; Luckin et al., 2022).

2.2. Hypotheses development

Drawing on Karaca et al. (2021), teachers' AI readiness consists of four components: cognition, ability, vision, and ethics. Aligned with the current research context, the component of cognition refers to teachers' cognitive readiness, involving knowledge about the functions of AI, importance of AI for education, and relationships between AI and human teachers. The ability component is related to teachers' competence and skills in the use of AI for teaching, for instance, selecting AI technologies appropriate for different activities, and designing and refining AI pedagogy for better education. Vision is concerned with teachers' perceptions of strengths and limitations of AI for education and insights into opportunities and challenges involved. Compared with cognitive readiness which emphasizes teachers' knowledge of AI for education, vision is more focused on their ability to envision and explore the potential and boundaries of AI in education. The component of ethics refers to teachers’ compliance with ethical and legal norms and regulations related to the use of AI for education.

Even though the ethical issue has been attracting increasing attention (Hagendorff, 2020), limited is known about what may contribute to individuals' ethical knowledge and practice. While external regulations such as ethical guidelines are important, research suggests that they may be insufficient in influencing individuals' ethical decision-making (Hagendorff, 2020; Mittelstadt, 2019). Therefore, it is essential to explore whether teachers' internal factors, such as their cognition, ability, and vision in the use of AI, may predict their ethical use of AI in education. By considering both external regulations and internal factors that influence ethical decision-making, we can take a more comprehensive approach to promoting teachers’ ethical use of AI in education. Thus, the following hypotheses are proposed:

Teachers’ cognition (H1a), ability (H1b), and vision (H1c) in the educational use of AI are positively associated with their ethical use of AI.

As AI is likely to cause great changes to people's lives in the foreseeable future, its expansion in the field of education appears to be threatening for some educators across different levels of schools (Chounta et al., 2022; Walia & Kumar, 2022). Teachers' perceptions of AI threats are manifested in different ways, such as unsafe feelings about their identities in education, job insecurity, and disruptions to their conventional work (Mirbabaie et al., 2022). However, those who demonstrate higher readiness for the use of AI may embrace AI with more confidence and are likely to adopt innovative behaviors, such as risk-taking, experimentation with new pedagogy, and problem-solving (Jöhnk et al., 2021; Microsoft, 2020). In this sense, AI-enhanced innovation goes beyond introducing more advanced technologies into more classrooms (Popenici & Kerr, 2017). Teachers with high AI readiness can reinvent approaches to teaching in order to better prepare students for the future society (Schleicher, 2015). Therefore, the following hypotheses are developed:

Teachers’ cognition (H2a), ability (H2b), vision (H2c), and ethics (H2d) in the use of AI are positively associated with their AI-enhanced innovation.

Teachers’ cognition (H3a), ability (H3b), vision (H3c), and ethics (H3d) in the use of AI are negatively associated with their perceived threats from AI.

AI can free teachers from monotonous administrative and teaching work and assist them to focus on innovative work such as developing students' higher-order thinking skills (Belpaeme et al., 2018). Teachers can add humanity to the deployment of AI in education by designing and implementing AI pedagogy and offering social and emotional care (Felix, 2020). The symbiotic interaction between teachers and AI may form a positive complementarity and strengthen teachers' innovation at work, eventually leading to increased job satisfaction (Nazareno & Schiff, 2021), which refers to a positive emotional state attained from one's appraisal of his/her job performance (Locke, 1976).

Although AI is expected to revolutionize learning and teaching in education (Zawacki-Richter et al., 2019), many educators are apprehensive about the possibility of being replaced or having their skills become obsolete (Celik et al., 2022; Chounta et al., 2022). This sense of threat posed by AI may discourage educators from taking risks, ultimately hampering their ability to innovate in teaching with AI (Jöhnk et al., 2021; Kim & Kim, 2022). In addition, perceived threats from AI may lead to a sense of job insecurity and cause anxiety in teachers (Bhargava et al., 2021), thereby eventually undermining their job satisfaction. Informed by the analyses above, the following hypotheses are proposed:

H4

AI-enhanced innovation is positively associated with teachers' job satisfaction.

H5a

Perceived threats from AI are negatively associated with AI-enhanced innovation.

H5b

Perceived threats from AI are negatively associated with teachers' job satisfaction.

Overall, the hypotheses of this study are visualized in Fig. 1.

3. Methodology

3.1. Participants and research contexts

The participants of this study were recruited from 19 cities in eastern China. To respond to the call from MOE of China to modernize school education using AI (Yan & Yang, 2021), the educational bureaus of these cities have been working to deploy AI for education following the top-down approach by assisting primary and secondary schools to collaborate with technology vendors or help them develop AI applications based on their talent resources.

This study utilized convenience sampling for the participant selection. With the assistance of the educational bureaus of these cities, primary school teachers who were involved in this scheme were approached through an online survey platform, as primary schools normally face less pressure from entrance examinations than secondary schools and thus were active in adopting new technologies for improving and diversifying education. This phenomenon is supported by Celik et al. (2022) who found in their review of teachers’ use of AI that primary education was the domain where AI was most frequently used by teachers. Moreover, teachers with prior experience using AI may possess a strong appreciation for the importance of AI readiness and its impact on their professional responsibilities.

As the researchers had no access to information about the total number of the teachers who were invited to participate, the response rate could not be calculated. After excluding invalid responses, the present study retained valid responses from 3164 out of 3950 participants. These valid responses included 1264 teachers from downtown areas, 943 teachers from town areas, and 957 teachers from village areas. Among the participants, 432 were males and 2732 were females, with an average age of 36.82 (SD = 8.10). The participating teachers taught courses ranging from Year 1 to Year 6, including literacy, mathematics, English as a foreign language (EFL), chemistry, music, and so on. Those teaching EFL made up the majority of the participants (N = 2236), followed by those teaching literacy (N = 392), mathematics (N = 323), and other courses (N = 213). The popularity of AI in EFL is probably due to the widespread adoption of AI-powered language learning applications thanks to the advances in natural language processing technologies (Wang et al., 2023; Randall, 2019). AI technologies such as chatbots have also been adopted by teachers to teach literacy and Chinese vocabulary (Chen et al., 2020). As for mathematics learning, many primary school teachers start to use intelligent tutoring systems for automated grading (Hwang & Tu, 2021).

3.2. Instrumentation

Besides the items gathering participants' demographic information, the 31-item survey instrument comprised seven variables, including the four variables of AI readiness (cognition, ability, vision, and ethics), AI-enhanced innovation, perceived threats from AI, and job satisfaction (see Appendix A). The items were rated on a five-point Likert scale where 1 indicated “strongly disagree” and 5 “strongly agree”. The four variables of AI readiness were adapted from Karaca et al. (2021). Five items were used to represent the concept of cognition in the use of AI, for instance, “I understand how AI technologies are trained and function in education.” Six items represented the ability to use AI for teaching, for instance, “I can optimize and reorganize the teaching process with the help of AI technologies.” There were three items indicating the concept of vision in the use of AI for teaching, for example, “I foresee the opportunities and challenges that AI technologies entail for education.” There were four items measuring ethics in the educational use of AI, for example, “I use the data of teachers and students generated by AI systems following legal and ethical norms.” The Cronbach's alpha values of cognition, ability, vision, and ethics in this study were 0.93, 0.97, 0.90, and 0.93, respectively (see Table 1 in Section 4).

 

Appendix A. Survey items for AI-readiness

Constructs		Items	Sources
AI-readiness	Cognition	(CO1) I clearly understand the new role of teachers in the era of AI. (CO2) I can effectively balance the relationship between teachers and AI technologies. (CO3) I understand how AI technologies are trained and function in education. (CO4) I can distinguish the functions and features of different AI tools and applications. (CO5) I understand the importance of utilizing AI technologies for data collection, analysis, evaluation, and security in education in the era of AI.	Karaca et al. (2021)
	Ability	(AB1) I can effectively integrate AI technologies into my classroom routines. (AB2) I can design different teaching approaches based on different functions of AI technologies. (AB3) I can rationally use AI technologies to solve problems discovered during the teaching process. (AB4) Based on the visual and real-time feedback provided by AI technologies, I can improve my teaching in the next step. (AB5) I can optimize and reorganize the teaching process with the help of AI technologies. (AB6) I can effectively discuss, share, and collaborate with other teachers on the use of AI technologies to jointly design high-quality teaching solutions.
	Vision	(VI1) I understand the strengths and limitations of AI technologies. (VI2) I have my own unique thinking and views on how to improve and use AI technologies for education. (VI3) I foresee the opportunities and challenges that AI technologies entail for education.
	Ethics	(ET1) I understand the digital ethics that teachers should possess in the era of AI. (ET2) I understand the ethical obligations and responsibilities teachers need to assume in the process of using AI technologies. (ET3) I know how to keep personal information safe when using AI technologies. (ET4) I use the data of teachers and students generated by AI systems following legal and ethical norms.
Perceived threats from AI		(PT1) I feel that AI technologies could weaken the importance of teachers in education. (PT2) I feel that the use of AI technologies has reduced the frequency of face-to-face communication with colleagues and students. (PT3) Students’ overreliance on the learning guidance provided by AI technologies may undermine the relationship between teachers and students. (PT4) I think that frequent use of AI technologies to assist teaching and learning may lead to inertia, which may reduce the thinking and decision-making abilities of teachers and students. (PT5) In my opinion, overuse of AI technologies may reduce the necessity of human teachers in the classroom, rendering it difficult for teachers to pass on correct values ​​to students.	Mirbabaie et al. (2022)
AI-enhanced innovation		(INN1) AI technologies enable me to accomplish tasks that were previously difficult to do without them. (INN2) AI technologies allow me to experiment with innovative pedagogy. (INN3) AI technologies enable me to organize teaching innovatively.	Popenici and Kerr (2017)
Job satisfaction		(JS1) In most ways, my job is close to my ideal. (JS2) The current condition of my job is excellent. (JS3) I am satisfied with my job. (JS4) I feel a sense of pride in my job. (JS5) My job is enjoyable.	Ragu-Nathan et al. (2008)