Maneuvering through the stormy seas of digital transformation: the impact of empowering leadership on the AI readiness of enterprises

ABSTRACT

Digital transformation exposes enterprises of altering business strategies and adapt technological advancements such as artificial intelligence (AI). AI is expected to fundamentally transform the future of work, however, associated changes cause resistance behaviour of employees and hindered AI readiness, i.e., missing preparedness for the implementation. This study examines whether leadership reduces resistance and how it contributes to the AI readiness of enterprises. Expert interviews indicate that empowering leadership, precisely autonomy and development support, is favourable to maneuver AI-induced change. Nevertheless, the quantitative evaluation on the impact of empowering leadership shows neither a significant effect on resistance to change nor on AI readiness. We discuss that employees strive for consistency and leaders who provide stable environments. Researchers understand which factors implicate AI readiness of employees and how leadership and resistance to change contribute. Practitioners comprehend which leadership attributes are fruitful for organizational alignments and how leaders generate appropriate readiness.

KEYWORDS:

• Digital transformation
• artificial intelligence
• leadership
• readiness
• resistance to change
• enterprises

1. Introduction

In today’s business world, organisations are obliged to face constantly changing technological advancements inevitably reshaping business models, internal processes, products or services. Thereby, the current wave of the digital transformation holds various opportunities but also confronts enterprises with certain challenges (Heavin & Power, 2018; Hess et al., 2016; Wessel et al., 2020; Frick & Marx, 2021), ranging from altered customer expectations to the development of new business strategies (Vial, 2019). However, they also include structural changes, such as leadership and employee roles, as well as organisational barriers, for example, resistance to change (Vial, 2019).

Digital transformation is related to the pressing need of organisational alignments and the demand to change business strategies for being able to contend with competitors and fulfill customer demands (Vial, 2019). This metamorphosis can be defined as “a process that aims to improve an entity by triggering significant changes to its properties through combinations of information, computing, communication, and connectivity technologies’ (Vial, 2019). Digital transformation thus addresses changes associated with the application and integration of new information technology in existing organisational structures (Hess et al., 2016; Verhoef et al., 2019). Furthermore, it is also concerned with the enhancement of customer value propositions using disruptive technologies (Nambisan et al., 2019). One technology that is expected to have a major impact on adapting organisations (Benbya & Leidner, 2018; Yan et al., 2018; Yang & Siau, 2018a) and their future of work (Grønsund & Aanestad, 2020; Wang & Siau, 2019) is artificial intelligence (AI).

The term AI covers a wide range of technologies with self-learning abilities that are capable to achieve above human performance (Batin et al., 2017; Coombs et al., 2020). AI is applied to support employees in the decision-making process (Brachten et al., 2020; Mirbabaie et al., 2020) or facilitate strategic decisions on an organisational level (Aversa et al., 2018), thus has a strong economic impact that frees humans from unwanted or repetitive tasks (Yang & Siau, 2018b). Due to its unique characteristics, it is expected that AI will massively change the work in enterprises (Grønsund & Aanestad, 2020; Zaza et al., 2019) eventually replacing or at least altering certain jobs (Coombs et al., 2020; Duan et al., 2019; Erdélyi & Goldsmith, 2018). There are varying perceptions among employees what AI is really capable of, influenced by science fiction and media, whereas the sociotechnical context in which AI is deployed is often ignored (Aleksander, 2017; Borges et al., 2020; Johnson & Verdicchio, 2017). The resulting discomfort might lead to enterprises and their employees not being ready to introduce AI; thus, the upcoming change is likely to fail.

Research indicates that leadership and its interplay with employees are crucial for change processes and innovation (Ahmad et al., 2020; Stefanou, 2001) but especially for digital transformation (Baptista et al., 2020; Heavin & Power, 2018; Prince, 2017; Vial, 2019). In the context of AI, leadership can be the decisive factor between success and failure (Alsheiabni et al., 2019; Lichtenthaler, 2020). A leader is an individual influencing other people to embrace enthusiastic, emotional, and physical energy focused on the effort to achieve an organisational goal (Winston & Patterson, 2006). Leaders are certainly able to positively influence emerging resistance behaviour. The success of a change process is said to be pertained to the manner in which the particular change is managed (Armenakis et al., 1993; Oreg & Berson, 2019; Oreg et al., 2011).

There is an urgent demand to conduct deeper research focusing on AI in the context of business value and associated changes (Coombs et al., 2020). The perception that AI will spiral out of control and making humans irrelevant might be unfounded but yet exists (Coombs et al., 2020; Johnson & Verdicchio, 2017). Nevertheless, organisations need to invest in AI and introduce applications to remain competitive (Benbya & Leidner, 2018; Schuetzler et al., 2018; Yan et al., 2018). However, upcoming deployments of AI are repeatedly accompanied by anxieties of employees leading to hindered AI readiness of organisations. Following Alsheibani et al. (2018), we define AI readiness as the preparedness of an enterprise and its employees for the implementation of a change concerning technologies related to the concept of artificial intelligence. Organisations that experience a high-level of readiness for change possess employees that tend to be highly invested in the change effort (Abdel-Ghany, 2014). Failed change projects are related to the inability of organisations to create readiness for change (Jones et al., 2005) where employees’ readiness needs to be addressed prior to the implementation (Abdel-Ghany, 2014). Enterprises introducing new complex digital innovations, such as AI, are required to focus on mediated perspectives rather than the technological implementation (Dewi et al., 2018; Yang & Siau, 2018b). Organisational leaders seize influential value regarding the employees’ motivation, job satisfaction as well as work climate for encouraging a change process (Damanpour & Schneider, 2009; Oreg & Berson, 2015, 2019). We thus argue that leadership has a strong impact on the AI readiness of enterprises and resistance behaviours of employees, therefore influencing upcoming change processes. Accordingly, we pose the following research questions:

RQ1: Which leadership attributes are favourable for an AI-induced change process?

RQ2: How do leadership attributes and resistance to change factors contribute to the AI readiness of employees in enterprises?

A mixed-method approach was followed for answering our research questions. For adressing our first research question, we identified leadership attributes by conducting 9 interviews with leaders (N = 4) and their employees (N = 5). To answer the second research question, and related to the findings from the interviews, we developed our hypotheses and evaluated them quantitively using an online survey with N = 128 participants.

This paper contributes to theory and practice by extending our understanding of leadership and its impact on resistance behaviours as well as the AI readiness of organisations and their employees. Researchers will find the insights fruitful in understanding which factors implicate AI readiness of employees. Furthermore, what leadership attributes and resistance to change factors contribute to this effect. Practitioners, such as managers and executives, will be able to comprehend which leadership attributes are particularly suitable for organisational alignments within the digital transformation. Managers supervising employees, for instance, frontline or functional managers, will assess the impact of leadership behaviours. This will aid in tackling managerial boundaries as our results provide evidence on how leaders need to deal with employees before an upcoming change process, thus generating the highest possible AI readiness of enterprises. Organisations might use our insights to prepare their management by coaching leadership attributes that are favourable for AI-induced changes. We believe this study will be valuable to scholars and professionals equally for understanding and overcoming obstacles when planning and initiating an AI-induced change processes. Our results prepare leaders and enterprises to create the best possible climate for an upcoming AI change process. Hence, this article extends the information systems (IS) literature by broadening our knowledge on leadership and resistance behaviour by examining its relevance for AI.

This paper is structured as follows: We describe the related work on leadership and its relation to resistance to change, followed by the research design including the hypotheses development. Afterwards, we explain our data collection and methods as well as the analyses and present our findings of the expert interviews and the online study. Finally, we discuss our key findings and their contributions to IS theory and practice.

2. Related work

Digital transformation and the emergence of new technologies, such as AI, is constantly changing the working environment in enterprises. In these stormy seas for employees, leadership is an indispensable component for successfully preparing, initiating and implementing a change.

2.1. Leadership and resistance to change

Leadership can be described as a process of directing other’s behaviour in a particular direction to achieve a certain objective (Kumar Sharma & Shilpa Jain, 2013). A leader is understood as ‘one or more people who selects, equips, trains, and influences one or more follower(s) who have diverse gifts, abilities, and skills and focuses the follower(s) to the organization’s mission and objectives causing the follower(s) to willingly and enthusiastically expend spiritual, emotional, and physical energy in a concerted-coordinated effort to achieve the organisational mission and objectives’ (Winston & Patterson, 2006). Successful leaders ensure that the enterprise develops a digital mindset while positively responding to new digital technologies and disruptive changes (Benlian & Haffke, 2016; Hansen et al., 2011).

Leadership has been researched in the context of digital transformation including arising challenges and adjustments (By, 2020; Heavin & Power, 2018; Montealegre et al., 2019; Prince, 2017), where certain characteristics were found to be effective (Khan, 2016; Larjovuori et al., 2018; Montealegre et al., 2019; Prince, 2017; Richter & Wagner, 2014; Wagner, 2016). For example, Prince (2017) shaped the wording ‘leadership 4.0’ which is based on the assumption that ‘leaders see the future, explore the future, and return to train others’. Moreover, Amundsen and Martinsen (2014) introduced the concept of empowering leadership including specific attributes incorporating the enhancement of an individual’s motivation through the delegation of responsibility and authority. As digital transformation is constantly changing the working environment, it becomes of certain interest to support self-leadership skills of employees (Amundsen & Martinsen, 2014, 2015; Carte & Becker, 2016). Leaders who are capable of creating a trustful environment and inspiring employees reduce resistance to upcoming changes (Oreg, 2006).

Resistance to change is defined as forces that counteract changes at work (Laumer, 2011). Within the IS literature, there are varying perceptions of what constitutes to resistance behaviour (Laumer & Eckard, 2010). There have been several approaches in investigating resistance to specific technologies and related attitudes (Alkraiji et al., 2016; Beare et al., 2020; Hossain et al., 2019; Silva et al., 2016). Factors contributing to resistance to change are, for example, new responsibilities, lack of trust and conceptual training, increased job complexity, and perceived threat to status and power (Hee-Woong & Kankanhalli, 2009; Joshi, 1991; Krovi, 1993; Laumer, 2011; Meissonier & Houzé, 2010). More IS research derived corresponding literature reviews (Ali et al., 2016; Hee-Woong & Kankanhalli, 2009; Laumer, 2011) and included a psychological and social-psychological lens (Lin et al., 2018). However, recent studies do not acknowledge the unique characteristics of AI that might alter the findings. Thus, the question of why individuals reject certain changes still remains less explained (Laumer, 2011).

Deeper research is required as the ongoing digital transformation, including the emergence of AI, forms new ways of working thus demanding new ways of leading (Hesse, 2018; Oberer & Erkollar, 2018; Sawy et al., 2016). AI and its multi-faceted capabilities are relatively new for most organisations including their employees. There is only a vague picture of related consequences associated with the deployment of AI (Kühl et al., 2019), which creates resistance behaviour of employees. However, this behaviour can be mitigated or eliminated by a leader.

2.2. AI in enterprises

The digital transformation focuses on several technologies (e.g., Alshareef et al., 2019; Seebacher & Ronny, 2019; Shahbaz et al., 2019) but AI receives an ever-increasing attention in research and practice.

AI is not a single technology, but rather a group of concepts that is constantly evolving with potential applications in almost every field (Barredo Arrieta et al., 2020). There is no unified definition, but AI can be considered as ‘the ability of a machine to perform cognitive functions that we associate with human minds, such as perceiving, reasoning, learning, interacting with the environment, problem-solving, decision-making, and even demonstrating creativity’ (Rai et al., 2019). AI is believed to fundamentally change the working environment as well as the behaviour of employees (Bednar & Welch, 2019). It is said that AI will frame the future of business across industries (Wang & Siau, 2019). Enterprises are sensing economic potential, sustained competitive advantage and maximising the market share (Benbya & Leidner, 2018; Yan et al., 2018).

AI comes along in various shapes. There are autonomous applications in organisations (e.g., N. Frick et al., 2019), services enriching existing IT-systems (e.g., Frick et al., 2019, 2020) or agents using AI capabilities to interact with employees (e.g., Brachten et al., 2020; Mirbabaie et al., 2020). Research and practice forecast a forthcoming of an AI revolution in which a wide range of jobs will be automated or replaced and the capabilities of AI tend to exceed human performance (Grace et al., 2018; Makridakis, 2017). A recent study of N. Frick et al. (2019) demonstrated that AI can be integrated into internal workflows of organisations. The authors showed that results generated by the AI are superior compared to human efforts and massively accelerate processes as well as free humans from unwanted duties. However, this might lead to employees becoming obsolete probably losing their jobs. Furthermore, the authors argue that an introduction should be carried jointly involving management and employees as well as encouraging a culture of open communication. Another study of Pessach et al. (2020) analysed an AI-based decision support tool for the recruitment and placement of professionals. The results indicate a high degree of precision compared to a human selection eventually replacing employees to save resources. They further argue that an increased return of investment is achievable and overall recruitment costs can be lowered.

The examples illustrate the raising concerns of employees’ that the introduction of AI in enterprises might alter jobs or even replace them entirely (N. Frick et al., 2019; Grace et al., 2018; Makridakis, 2017), something which is highly discussed in the media (Willcocks, 2020). However, the ongoing digital transformation poses numerous benefits for organisations, such as maintaining a competitive advantage, which cannot be overlooked (Hess et al., 2016; Wimelius et al., 2020). Nevertheless, the potential impact of AI, as highly disruptive concept, on employees in enterprises cannot be neglected. Organisations planning to introduce AI need to ensure that their employees are ready for the upcoming change process. Literature indicates that failed change projects are related to the missing preparedness of organisations (Jones et al., 2005) with employee’s readiness as an essential factor that needs to be addressed prior to the upcoming change (Abdel-Ghany, 2014). IS Scholars showed that the management is of particular relevance when AI is planned to be applied in organisations (Alsheiabni et al., 2019; Lichtenthaler, 2020). The employees’ attitudes towards a change are relevant in understanding the organisational change process (Caldwell et al., 2004; Fugate et al., 2008; Oreg, 2006; Rafferty & Griffin, 2006). Likewise, as leadership is known to be perceived subjectively (Crawford & Kelder, 2019), and as environmental events are said to be perceived subjectively, it is necessary to focus on individual differences with regard to the attitude towards a change. Following recent research (e.g., Baptista et al., 2020; Coombs et al., 2020; Heavin & Power, 2018; Vial, 2019), we need to broaden our understanding of leadership and resistance to change with the increasing occurrence of AI in organisations. To our knowledge, there is no empirical evidence of whether leadership contributes to an upcoming AI change. In particular, which leadership attributes are favourable for the AI readiness of employees in organisations, further, its relation to resistance to change.

3. Research design

In order to examine how leadership attributes are favourable for AI change processes and further how these attributes and resistance to change contribute to the AI readiness of employees in enterprises, we selected a mixed-method approach. This design strategy provides a powerful mechanism for IS researchers in dealing with evolving situations and complex advancements while generating contributions for theory and practice (Venkatesh et al., 2013). Mixed-methods are fruitful as they simultaneously might address confirmatory and exploratory issues, provide greater insights compared to separate methods and pose the opportunity to analyse divergent and/or complementary findings (Teddlie & Tashakkori, 2003, 2009). Our exploratory sequential procedure combining qualitative and quantitative research aims at validating whether assumptions based on a small sample size can be generalised in a greater domain (Creswell & Creswell, 2018; Kelle, 2006). We use qualitative research to explain natural phenomena, thus identifying core issues and developing hypotheses (Kelle, 2006) while the subsequent quantitative step objectifies validating derived hypotheses with a larger population (Creswell & Creswell, 2018). In terms of sample size, Creswell and Creswell (2018) advise including between three and ten individuals for the qualitative part while Morse (1994) argues to acquire more than 5 participants. Within the quantitative method, Onwuegbuzie and Leech (2005) favour using at least 64 participants for one-tailed, and 82 participants for two-tailed hypothesis. Following the outlined recommendations, we conducted 9 expert interviews with leaders (N = 4) and employees (N = 5) to identify leadership attributes which are suitable to manoeuvre an AI-induced change process. Based on the findings, and grounded by theoretical foundations, we derived our hypotheses for AI readiness which were quantitatively validated via an online survey (N = 128).

3.1. Interviews

As there is little empirical evidence on which leadership attributes are relevant in the context of an AI-induced change within organisations, we used semi-structured interviews as the most appropriate method to obtain knowledge from experts (Kvale & Brinkmann, 2009). To prevent purely unilateral viewpoint, we conducted nine interviews with employees and their direct leaders from different organisations across industries. Leaders might portray themselves differently than they are actually perceived by their employees. Thus, this approach ensures gaining a holistic picture considering the perspectives of both, leaders and employees. We further assured that experts already experienced a technology-induced change process. The experts were between 25 and 50 years (M = 34.5; SD = 8.23) with 3 female and 6 male subjects. We cover three industrial sectors with six different positions.

The open interview technique was used as the most applicable way for retrieving valuable data and providing sufficient room for experts to elaborate on their subjective opinions (Meuser & Nagel, 2009). We created a prefixed guideline with central questions considering literature from the method of expert interviews and the guiding concept of leadership (e.g., Amundsen & Martinsen, 2014, 2015; Carte & Becker, 2016). The interview recordings were paraphrased following the recommendations in the qualitative assessment of content analysis (Schilling, 2006). We thus reduced the data by removing unnecessary words to form short and concise sentences. We carefully listened to the interviews and paraphrased the content of the experts’ statements and further generalised and reduced the content to comprehend and interpret the meaning of the explanations. The data were coded following a deductive procedure since the categories are based on recent literature on leadership attributes (Mayring, 2015). Within the categorisation process, we aimed at identifying which attributes are relevant for a leader in AI-induced change processes. This research approach can be classified as a descriptive procedure since experts are describing the current situation and their experience (Bear & Knobe, 2016). An overview of our sample as well as the interview guide can be found in Appendix A (c.f. Tables A1 and Tables A2).

3.2. Hypotheses development

Leadership has been identified as key element for the implementations of innovations (Damanpour & Schneider, 2009) as well as encouraging employee’s change behaviours (Damanpour & Schneider, 2009; Oreg & Berson, 2015, 2019). The success of change processes is related to how a change is communicated, coordinated and managed by the leader (Armenakis et al., 1993; Oreg & Berson, 2011, 2019). Furthermore, leadership is vital for ensuring the readiness of employees for an upcoming change (Damanpour & Schneider, 2009; Oreg & Berson, 2015, 2019).

As outcome of our interviews, experts described a favourable leader for change processes as someone who provides a supportive environment in which employees can self-develop and self-realise. Furthermore, a leader was explained as an individual who takes care of employees and protects them from external negative influences.

“[…] who takes on employee responsibility regardless of the number of employees.¹” (Interviewee 9)

“[…] provides a suitable environment for employees to develop.” (Interviewee 6)

“[…] provides assistance with the self-realization of the employees.” (Interviewee 7)

“This requires a high degree of energy. Since there will be many setbacks. You have to be enthusiastic and to be able to pass this enthusiasm on to the employees.” (Interviewee 3)

The outlined attributes are associated to empowering leadership as they are addressing the enhancement of employee’s motivation through the delegation of responsibility and authority (Amundsen & Martinsen, 2014). Empowerment is an enabling process for fostering a person’s perception of self-efficacy (Lamm & Gordon, 2010). Leaders encourage employees to take independent actions and thus facilitate self-leadership. Besides transparent communication, this was highly encouraged by the experts.

“[…] should encourage employees to be open to new technologies.” (Interviewee 2)

“[…] communicate the benefit to each individual employee.” (Interviewee 9)

“[…] transparency and involving the team.” (Interviewee 5)

The failure of change implementation is linked to a lack of the intrapersonal preparedness of employees (Soumyaja et al., 2011), which is a result of the inability of enterprises to create readiness for change (Jones et al., 2005). Readiness is a key factor determining employees’ support for organisational change projects (Holt et al., 2007). Organisations who experience a high-level of readiness for change are said to possess employees that tend to be highly invested in the change effort (Abdel-Ghany, 2014). Based on the essential values of leadership regarding readiness for change (Amundsen & Martinsen, 2014), supported by the results of the qualitative assessment, we found autonomy support as well as development support as favourable leadership attributes for upcoming change processes and thus for the AI readiness of employees in organisations. We hypothesise:

H1a: The autonomy support of leadership is positively related to the employee’s emotional, intentional and cognitive AI readiness.

H1b: The development support of leadership is positively related to the employee’s emotional, intentional and cognitive AI readiness.

Change might result in non-adoption, rejection and resistance (Hirschheim & Newman, 1988; Keen, 1981; Laumer, 2011). The phenomenon of resistance to change describes the non-adoption behaviours (Eckhardt et al., 2009; Laumer, 2011) which involve a discordance of employees’ feelings, behaviours, and thoughts about the change (Oreg, 2006; Piderit, 2000). Non-adoption behaviours are a response to ongoing or present events which are perceived as inequitable, threat or stressful feeling (Laumer, 2011). Resistance to change is recognised as one of the key barriers to effective IT organisational implementation projects (Klaus & Blanton, 2010; Laumer, 2011), applying to any kind of change (Abdel-Ghany, 2014).

With regard to the performance of an enterprise and the success factors of projects, organisational leaders have been assigned an influential value regarding the employees’ motivation, job satisfaction as well as creating a trustful work environment and encouraging change processes (Damanpour & Schneider, 2009; Oreg & Berson, 2015, 2019). Leadership is crucial regarding the success of change processes across industries (Helming et al., 2019) and a key challenge for teams who are becoming virtual through digital transformation (Gallenkamp et al., 2011). Furthermore, leaders have been found to function as role model. The behaviour of a leader, including planning and coordination of changes, give rise to the performance strategies applied by employees (Zaccaro et al., 2001). More specifically, lack of faith in the organisation’s leadership has been strongly related to increased resistance to change attitudes and behaviours (Oreg, 2006). Hence, we hypothesise:

H2a: The autonomy support of leadership is negatively related to the employee’s routine seeking, emotional reaction, short-term focus and cognitive rigidity.

H2b: The development support of leadership is negatively related to the employee’s routine seeking, emotional reaction, short-term focus and cognitive rigidity.

Changes in enterprises caused by the digital transformation evoke a spectrum of responses (Lamm & Gordon, 2010), identifying resistance to change and the insufficient readiness of employees to be one of the main reasons why change projects fail projects (Klaus & Blanton, 2010; Laumer, 2011). Projects are likely to fail when enterprises are not sufficiently ready for an upcoming change (Jones et al., 2005). Furthermore, employee’s readiness is a decisive factor that needs to be considered prior the change process (Abdel-Ghany, 2014). When preparing AI readiness in organisations, we hypothesise a negative relation between employee’s dispositional resistance to change and employee’s readiness for AI-induced change. This means that employee’s tendencies to resist change will result in a lower AI readiness. We therefore pose:

H3a: Employee’s routine seeking is negatively related to the employee’s emotional, intentional and cognitive AI readiness.

H3b: Employee’s emotional reaction is negatively related to the employee’s emotional, intentional and cognitive AI readiness.

H3c: Employee’s short-term focus is negatively related to the employee’s emotional.

intentional and cognitive AI readiness.

H3d: Employee’s cognitive rigidity is negatively related to the employee’s emotional, intentional and cognitive AI readiness.

The effect of resistance to change on AI readiness is anticipated to be moderated by empowering leadership applied by the supervisor (Alsheiabni et al., 2019; Alsheibani et al., 2018). This is grounded on the value of leadership within the successful implementation of change as well as the impact of leadership on the overall performance of an organisation. It can be assumed that leadership has the potential to influence the relationship between resistance to change and AI readiness by moderating the negative impact of resistance to change on AI readiness (Damanpour & Schneider, 2009). We therefore hypothesise:

H4: Empowering leadership moderates the correlation of resistance to change on AI readiness.

Our final hypotheses, as outlined in Figure 1, thus consider empowering leadership (autonomy support and development support), resistance to change (routine seeking, emotional reaction, short-term focus and cognitive rigidity) and AI readiness (emotional, intentional and cognitive).

Figure 1. Constructs and hypotheses

4. Data collection and analysis

To analyse how leadership and resistance factors contribute to the AI readiness of employees, we followed both a qualitative and quantitative approach. The conducted interviews led to the assumption that empowering leadership is fruitful for AI changes in organisations. We evaluated the findings as well as resistance to change factors using a quantitative online survey.

The online study started with a briefing about the content and purpose of the study and provided with information about the participants’ rights and the anonymous collection of data. Afterwards, the participants were asked to answer questions regarding their demographic data followed by questions on the specific constructs, i.e., empowering leadership, resistance to change and AI readiness. As last question, completed by a debriefing, the participants were asked whether they had experienced with an AI in organisations and associated change processes.

Subjects were recruited via business social networks such as LinkedIn or Xing according to certain criteria: Individuals needed to work in an organisation whereas students, interns, trainees were excluded. Furthermore, participants needed experience with a change process in their current or former company. Details on the online study and adapted items can be found in the appendix (cf. Table A3).

4.1. Measures

Our developed hypotheses were validated using a quantitative online questionnaire. Therefore, we adopted and modified constructs focusing on the context of our research. We used previously validated instruments to ensure high accuracy of the measures.

Empowering leadership, as identified within the interviews, was adapted from Amundsen and Martinsen (2014) who conceptualised valid and reliable scale determining the construct of empowering leadership. The Empowering Leadership Style Scale assesses key behaviours of leaders by using a two-dimensional 18-item instrument from Amundsen and Martinsen (2014). Example items are ‘Employees get an insight into the way leaders organise their working days’ and ‘Supervisors let employees see how he/she organises his/her work’. In our study, all scales demonstrate high reliability with Cronbach’s α = .91.

The level of resistance to change according to Oreg (2003) was used to measure the individual’s attitude towards change. The scales consist of 18-items covering a multidimensional view of change. Additionally, the Resistance to Change scale is validated as a reliable assessment for people’s untailored dispositional reactions towards change regardless of their position at work and is further not tailored to any specific type of change (Oreg, 2003). Example items are as ‘I generally consider changes to be a negative thing’ and ‘When things don’t go according to plans, it stresses me out’. The Resistance to Change scale had a high reliability with Cronbach’s α = .83.

The AI readiness of employees in organisations was determined following Bouckenooghe et al. (2009). The Organisational Change Questionnaire–Climate of Change, Processes, and Readiness; (OCQ–C, P, R) was applied to predict employee’s change behaviours and reactions. The subscale of the OCQ-C, P, R assess the readiness for change as a tridimensional attitude (Bouckenooghe et al., 2009), which allows a thorough interpretation of the phenomenon. Thereby, three factors are covered: emotional readiness, cognitive readiness for change and intentional readiness for change. Exemplary items are ‘I have a good feeling about the change project’, ‘The change will improve work’ and ‘I am willing to make a significant contribution to the change’. The scales show reliability with Cronbach’s α = .76.

All items in the study were slightly rephrased to transfer them into the context of our research. We used a 5-point Likert scale ranging from ‘not at all’ to ‘very much’ for each item to ensure consistency. In addition, we measured the age, gender, education of participants as control variables.

4.2. Descriptive statistics

Overall, 243 participants took part in the survey. Unfinished responses or suspicious responses as well as anomalies lead to exclusion of the record. The final dataset consisted of N = 128 participants. The participants were between 20 and 67 years (M = 40.86; SD = 12.13), 89 of them female (69%) and 39 male (31%). All participants work in offices located in Germany. Most of the participants (68 of 128) had an academic degree with 18 (14%) Bachelor graduates, 39 (31%) Master graduates and 11 (9%) doctorates. Sixty participants (46%) had a high school degree or equivalent. Overall, 58 subjects (36%) declared themselves to be in a managerial position and, finally, 39 participants (31%) indicated to have already experienced an AI-induced change process at their current or former employer.

5. Results

For validating our hypotheses, we assessed the correlation coefficient of the constructs using the Pearson correlation coefficient (PCC). The PCC is a reliable and widely accepted statistical metric which allows to measure the strength of a linear relationship between two variables (Zhou et al., 2016). Furthermore, the moderating effects of empowering leadership were calculated via regression analysis.

5.1. Empowering leadership and AI readiness

For validating whether empowering leadership is linked to the AI readiness of employees in organisations, each hypothesis was evaluated using Pearson correlations as displayed in Table 1. The results indicate that none of the correlations were found significant. Cognitive readiness displayed a small negative correlation with both autonomy support and development support. Whereas emotional readiness demonstrated a small positive correlation with both factors of autonomy support and development support. Intentional readiness was the only readiness factors that correlated positively with autonomy support and at the same time correlated negatively with development support.

Table 1. Pearson correlations for empowering leadership on AI readiness

	Autonomy support	Development support
Emotional readiness	.015	.046
Intentional readiness	.034	−.056
Cognitive readiness	−.054	−.046

5.2. Empowering leadership and resistance to change

The interaction between empowering leadership and resistance to change was examined calculating the Pearson correlations. Even though none of the correlations classified as significant, negative correlations were found between emotional reaction and autonomy and development support. Another negative correlation was identified between development support and routine seeking. The final values can be found in Table 2.

Table 2. Pearson correlations for empowering leadership on resistance to change

	Autonomy support	Development support
Routine seeking	.005	−.054
Emotional reaction	−.204	−.152
Short-term focus	.129	.271
Cognitive rigidity	.271	.154

5.3. Resistance to change and AI readiness

For testing the relationship between resistance to change and readiness for change, Pearson correlations were calculated. We found significant correlations of routine seeking on emotional readiness, routine seeking on intentional readiness and emotional reaction on intentional readiness. In addition, a linear regression analysis was performed indicating that the degree of empowering leadership accounts for 0.0 % of the variance in the emotional readiness in this person (R² = .000, F(1,68) = 0.001, p = .976). This indicates no direct effect between these two variables. Table 3 outlines the results.

Table 3. Pearson correlations of resistance to change on AI readiness

	Emotional readiness	Intentional readiness	Cognitive readiness
Routine seeking	−.197*	−.242**	.055
Emotional reaction	−.079	−.205*	.103
Short-term focus	−.054	−.163	.125
Cognitive rigidity	.123	.092	−.004

**. Correlation is significant at 0.01 level (2-tailed); *: Correlation is significant at 0.05 level (2-tailed).

5.4. Moderating effect of empowering leadership

To assess the interaction of resistance to change and empowering leadership as main effects on the AI readiness of employees, the interaction was compared against a baseline model of controls. The results of the regression analysis indicate that resistance to change and empowering leadership account for 2.6 % of the variance in the AI readiness of an employee (R² = .026, F(3,66) = .587, p = .625). Indicating that the level of experienced resistance to change and the performed empowering leadership did not significantly predict the AI readiness level of an employee. The results for the regression are depicted in Table 4.

Table 4. Regression results for the hypothesised moderation

Variable	B	SD	$\beta$	t	p
AI readiness	26.464	.687	-	38.531	.000
Moderation	−.003	.003	−.148	−1.215	.229
Empowering leadership	−.027	.068	−.071	−.399	.692
Resistance to change	.023	.052	.078	.435	.665

6. Discussion and implications

Digital transformation forces enterprises to change business strategies and empowers technological alignments (Hess et al., 2016; Verhoef et al., 2019). However, the application of AI as constantly evolving concept (Barredo Arrieta et al., 2020) confronts enterprises with certain challenges, e.g., resistance behaviour and missing AI readiness of employees. Further research is urgently needed to examine how leadership mitigate resistance for an upcoming AI change process. We thus followed a mixed-method approach for (a) identifying leadership attributes that are favourable for AI-induced changes and (b) evaluated their impact on resistance to change and AI readiness of employees in enterprises.

Surprisingly, we were neither able to find a significant correlation between empowering leadership and AI readiness nor between empowering leadership and resistance to change. We interpret this to the mean that empowering leadership is not favourable to enhance the AI readiness of employees in enterprises. The application of AI in organisations is believed to fundamentally change the working environment (Grønsund & Aanestad, 2020; Zaza et al., 2019) and is further likely to alter certain jobs or even replace them entirely (Coombs et al., 2020; Duan et al., 2019; Erdélyi & Goldsmith, 2018). Employees might simply be frightened of the AI-related consequences such as becoming superfluous and losing the job. Moreover, many employees do not know what AI is really capable of, which is additionally encouraged by science fiction and media (Aleksander, 2017; Borges et al., 2020; Johnson & Verdicchio, 2017). An empowering leader is thus not suitable for fostering AI readiness as this type seeks rapid changes and expects a high level of self-responsibility of employees (Lamm & Gordon, 2010). However, in the context of AI, employees might not be able to self-manage. We further interpret the missing correlation of empowering leadership and resistance to change to the mean that empowering leadership is not adequate for reducing employees’ resistance towards a change. Digital transformation, and especially AI, alters established structures and employees’ responsibilities in organisations (Grønsund & Aanestad, 2020; Vial, 2019; Zaza et al., 2019). However, changes initiated by disruptive concepts create discomfort among employees, for example, ambiguous impacts or outcomes, possibly leading to resistance behaviour (Kim, 2011; Metwally et al., 2019). This could mean that employees are more likely to prefer a leader which can provide a stable and reliable environment (Bolden & O’Regan, 2016; Picot et al., 2009). Furthermore, employees might even refuse their freedom for self-development and self-leadership since the change process itself is perceived as frightened.

The results further revealed a correlation between resistance to change and AI readiness. Explicitly, routine seeking is negatively correlated with emotional readiness and intentional readiness. Furthermore, emotional reaction negatively correlates with intentional readiness. We understand that employees who strive for consistency and are likewise more emotional, are less ready to adopt AI in their organisation. This goes in line with previous research (Eckhardt et al., 2009; Laumer, 2011) and explains that employees favour a stable and reliable environment (Bolden & O’Regan, 2016; Picot et al., 2009). This might also be due to a lack of knowledge and the fear that AI can achieve superior results compared to humans (Aleksander, 2017; N. Frick et al., 2019; Johnson & Verdicchio, 2017). Finally, our study did not identify a moderating effect of empowering leadership. This is not unexpected as we comprehend that empowering leadership does not correlate with resistance to change or AI readiness of employees in organisations.

The results of this research are manifold. To our knowledge, this study is the first to evaluate the intertwined relationship between leadership, resistance to change and the AI readiness of employees in enterprises. Hence, we provide unique insights into the phenomenon of AI readiness by adding to the IS literature the relationships of leadership as well as resistance to change. Although we have not been able to indicate that empowering leadership is favourable for improving AI readiness, we believe that leadership remains crucial and demands effective management strategies in organisations. We have motivated this research with the mismatch of employees’ expectations of AI and its actual capabilities. However, both, the time factor as well as the use of AI in private life, play a decisive role. The longer AI is used, and the more widespread it is, the higher the AI readiness in enterprises. In terms of practical implications, enterprises need to understand that AI, unlike other technologies in the context of digital transformation, has a special reputation that frightens employees. Enterprises need to carefully prepare their management for AI and related changes as here might also arise resistance behaviour or missing readiness. Only those leaders who are ready for an upcoming AI change can positively influence their employees.

Our research must be interpreted with certain limitations. First, the sample of our experts may not have been adequate to identify favourable leadership attributes for AI-induced change. The experts have all experienced a technology-driven change but not all of them were familiar with the introduction of AI in enterprises. In addition, most of our interviewees are rather young and work as consultants or managers. Both, the age and the more agile professions, may have misled us to identifying leadership attributes that more open to change. Second, we have used AI in general within our online study without narrowing down the scope to a specific application or domain. Since AI can be understood as a variety of technologies, this could have led the participants to having different types in mind.

Future research should examine the relationship between leadership and AI readiness on a multilevel and multi-factorial perspective. Therefore, studies need to address the impact of multiple leadership attributes as determinants of AI readiness of employees in enterprises. In addition, IS scholars might also be interested in examining AI readiness depending on the actual application or domain. We finally suggest a longitudinal field study in which an AI-induced change is accompanied by researchers and the AI readiness of employees is evaluated before and after the deployment process, with a special emphasis on the leader. This could shed a light onto practical conclusions from a real-world scenario.

7. Conclusion

Digital transformation is not a gust but a persistent storm that needs to be managed carefully. AI as integral component will change the working environment and habits of employees in enterprises. We exhibited relationships between leadership, resistance to change and AI readiness. It became clear that empowering leadership is not favourable when it comes to AI-induced changes. In fact, employees strive for consistency and a leader who provides a stable and reliable habitat. Thus, this research heavily impacts organisations and their managerial perspective including the strategy of dealing with employees. Freedom for self-development and self-leadership may not be suitable for certain change processes. In particular, managers and executives might need to rethink and alter their leadership styles in the context of digital transformation and AI. Enterprises might adjust their strategy towards providing employees with sufficient knowledge about AI and its current capabilities as well as fields of application, for instance, by offering training, seminars or workshops. Furthermore, it seems beneficial to involve affected individuals before an upcoming change process to counteract potential resistance behaviour. In summary, researchers need to focus on multiple leadership attributes while enterprises must prepare their management and employees for unavoidable AI implementations.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The excerpts from the German interviews have been translated into English for the readers convenience.

Appendix A.

Table A1. Interview guideline

Phase	Research goal	Questions
Introduction	Welcoming the interviewee and providing general information about the research and brief introduction to the topic.	-
Demographic data	Getting to know the expert.	a) Could you please introduce yourself? b) What is your current position in your organisation and what are your responsibilities? c) What is your professional background? d) What is your work experience? e) How many employees to you guide as manager? (managers only)
Definition and potential of artificial intelligence	Getting an understanding of the expert’s experience with AI and related changes.	a) What do you associate with the term ‘artificial intelligence’? Application areas b) What applications are you already using in you company? c) What applications are used by other companies that? d) How was your first contact with AI? e) Have you been involved in the development/decision making process to introduce AI? f) What changes or adaptations did the implementation of artificial intelligence entail? g) What are further applications of artificial intelligence (in the future)? Requirements h) Which tasks are likely to be executed by AI in your job? i) What requirements do AI applications have? j) Which organisational processes annoy you or are insufficient? k) Have you considered using AI in those areas? l) What are possible legal or ethical concerns?
Attributes of favourable leadership for AI-induced change	Identifying suitable leadership attributes for AI-induced change processes simultaneously contributing to AI readiness.	a) What do you associate with the term ‘leadership’? Requirements b) What expectations do you have of your leadership? c) What defines favourable leadership? d) Which attributes are relevant? e) Please select the most important attributes for a favourable leader (from the DISC Model). Experience f) What were your experiences with leadership in the past? g) Who is your current leader? h) Have you already gained your own experience as a leader? If so, can you explain them? If no, do you have ambitions to become a leader. i) How would you rate your current leader? j) How would you describe the personality of your leader? (e.g., extraverted (extroverted), sympatric (likeable), dependable (trustworthy), calm (calm), open to new experiences) k) What constitutes bad leadership? l) What is important to you when implementing new innovations?
Leadership and digital transformation	Identifying further leadership attributes that are relevant in the context of digital transformation.	How did the digital transformation change leadership? What leadership attributes are more preferable? Present and verify conceptual model
Finish	Debriefing of the interviewee and explanation of the research background. Possibility to give closing remarks.	-

Table A2. Descriptive and duration of the expert interviews

Id	Profession	Position	Gender	Age	Industry	Duration (min)
1	Senior Manager	Leader A	m	33	Financial services	60
2	Consultant	Employee A	m	25	Financial services	62
3	Director	Leader B	m	43	Capital projects & infrastructure	50
4	Manager	Employee B1	m	33	Capital projects & infrastructure	53
5	Consultant	Employee B2	m	28	Capital projects & infrastructure	50
6	Board member	Leader C	m	50	Tattoo & body	70
7	Head of department	Employee C	f	27	Tattoo & body	61
8	Partner	Leader D	f	40	Capital projects & infrastructure	57
9	Manager	Employee D	f	32	Capital projects & infrastructure	43

Table A3. Sequence of the online study

Group	Question	Type
Briefing
Demographic data	How old are you?	Numeric input
	What is your gender?	List selection
	What is your highest degree?	List selection
	What is your current working position?	List selection
	In which industry are you currently working?	List selection
	Size of your current company: (*Number of employees)	List selection
	Where is your current working location (country)?	List selection
Leadership experience	Are you currently in a managerial position?	Yes/No
	What management tasks do you mainly deal with?	List selection
Supervisor rating	What age group would you classify your current manager in?	List selection
	Please indicate the role that the manager you have selected is currently taking on in your work environment:	List selection
	In your opinion, would you describe your current manager as a favourable leader?	List selection
Empowering leadership style (adapted from Amundsen & Martinsen, 2014)	Please indicate to what extent these statements apply to you.	5-point Likert Scale (from “not at all” to “very much”).
Resistance to change (adapted from Oreg, 2003)	Please indicate to what extent these statements apply to you.	5-point Likert Scale (from “not at all” to “very much”).
Experience with AI	Have you already experienced an AI-induced change within your working environment?	Yes/No
	Please briefly describe your experienced AI-induced change.	Text input
AI Readiness (adapted from Bouckenooghe et al., 2009)	Please indicate to what extent these statements apply to you.	5-point Likert Scale (from “not at all” to “very much”).
Debriefing