AI Acceptance of Start-up Owners in the Entrepreneurial Ecosystem of Amsterdam

38  Download (0)

Full text


AI Acceptance of Start-up Owners in the Entrepreneurial Ecosystem of Amsterdam

Faculty of Economics and Business, University of Amsterdam Bachelor’s Thesis Entrepreneurship, Innovation and Creativity

Lisa Heynen

Student number: 12676144 Supervisor: Roel van der Voort Date: 29/06/2022

Word count main text: 6833

Word count abstract: 172



This document is written by Lisa Heynen who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no

sources other than those mentioned in the text and its references have been used in creating it.

UvA Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.



STATEMENT OF ORIGINALITY _______________________________________________ 1 ABSTRACT ________________________________________________________________ 3 1. INTRODUCTION _______________________________________________________ 4 1.1 Context _____________________________________________________________________ 4 1.2 Relevance ___________________________________________________________________ 4 1.2.1 Academic Relevance ______________________________________________________________ 4 1.2.2 Practical Relevance _______________________________________________________________ 5 1.3 Research question _____________________________________________________________ 6 2. LITERATURE REVIEW __________________________________________________ 6

2.1 Artificial Intelligence __________________________________________________________ 6 2.2 Start-up Owners ______________________________________________________________ 7 2.3 Start-ups in Ecosystem Amsterdam _______________________________________________ 8 2.4 Artificial Intelligence Acceptance Intention _______________________________________10 2.4.1 Performance Expectancy __________________________________________________________ 10 2.4.2 Openness ______________________________________________________________________ 11 2.4.3 Inconvenience __________________________________________________________________ 11 2.4.4 Uncertainty ____________________________________________________________________ 11 2.4.5 Attitude _______________________________________________________________________ 12

3. CONCEPTUAL MODEL ________________________________________________ 12 4. METHODOLOGY ______________________________________________________ 13 4.1 Empirical setting and data collection _____________________________________________13 4.2 Variables and measures _______________________________________________________14 5. RESULTS ____________________________________________________________ 15

5.1 Data cleaning, recoding, and correlations _________________________________________15 5.1.1 Reliability analysis ______________________________________________________________ 15 5.2 Descriptives and regression assumptions __________________________________________16 5.2.1 Linear relationship _______________________________________________________________ 17 5.2.2 Multivariate normality ____________________________________________________________ 18 5.2.3 Multicollinearity ________________________________________________________________ 19 5.2.4 Likert scale ____________________________________________________________________ 19 5.3 Linear regression ____________________________________________________________19 5.3.1 Hypotheses ____________________________________________________________________ 20 5.3.1 Alternative regressions ___________________________________________________________ 20

6. DISCUSSION _________________________________________________________ 21 6.1 Limitations and further research_________________________________________________22 6.2 Implications ________________________________________________________________23 6.3 Conclusion _________________________________________________________________24 7. REFERENCES ________________________________________________________ 24


Appendix A1 Demographic characteristics ___________________________________________29 Appendix A2 Scatterplots ________________________________________________________30 Appendix A3 Regression _________________________________________________________31 Appendix A4 Regression – Age > 40 ________________________________________________32 Appendix A5 Regression – Education >= Master’s degree _______________________________33 APPENDIX B – Survey questions ______________________________________________ 33


This research paper investigates the motivations behind the AI acceptance intention of start-up owners in the entrepreneurial ecosystem of Amsterdam. A simplified version of the ADIAE model by Upadhyay, Upadhyay and Dwivedi (2021) is used to create and test the hypotheses.

The theorized model has been validated with 69 usable responses. A multivariate regression generated results suggesting that performance expectancy and openness are positively related to AI acceptance intention. The rest of the predictors, uncertainty, inconvenience, and attitude did not confirm any significant relationships.

This thesis contributes to academic literature regarding the technological developments in Amsterdam. Policymakers can use the findings from this study to facilitate AI acceptance, businesses can use it to get more insights into the trends within the ecosystem, and employees may benefit from it to ensure a profitable position in the future labor force. Future researchers may replicate this research model or adapt it to their needs for further investigation. This study reveals the most important motivations behind the intention to accept AI of start-up entrepreneurs in Amsterdam.

Keywords: Entrepreneurship, Technology acceptance, Artificial intelligence, Amsterdam Metropolitan Area, Entrepreneurial ecosystems, Start-up entrepreneurs, Start-ups



1.1 Context

In the coming five years, the Netherlands will double its investments in technology and services that use Artificial Intelligence (AI) to 3.2 billion Euros (Business Insider, 2020). This will result in significant possibilities in various job sectors. However, with such disruptive developments and changes, some resistance will inevitably occur. Current employees are concerned that AI will take over their jobs and leave them unemployed. Per contra, the implementation of AI can ensure that employees will make their work more interesting by leaving the "boring and repetitive" tasks to the AI (Wilman, 2022). Therefore, the employees will have the ability to do things where humans are really needed. Still, resistance among society remains. To overcome this resistance, it is vital to locate the source of the resistance and establish where and why it is perhaps already accepted.

When organizations trust Artificial Intelligence more, they can relieve some of their employees' workload. They do not have to hire people for tasks that AI can already do. This paper examines the intention of start-up owners to accept AI. Furthermore, acceptance is an essential variable when looking at the implementation of AI. It can potentially help make new policies to overcome the resistance. Current literature is incomplete with respect to AI acceptance in start-ups, specifically in the entrepreneurial ecosystem of Amsterdam.

Amsterdam has a very innovative ecosystem, which is further explored in the literature review.

The start-ups of today are the large corporations of tomorrow. They can and will have a substantial impact on the labor market, as start-ups generate a significant number of new jobs (Terryn, 2021). Additionally, the type of jobs that are created is influenced by the implementation of AI, which is certainly affected by the intention to accept AI. Overall, this study will research the AI acceptance of start-up owners in Amsterdam.

1.2 Relevance

1.2.1 Academic Relevance

Existing literature explores many ways AI influences organizations, as it has become increasingly more abundant how important it is. This thesis will contribute to the existing knowledge about AI to keep the information up to date and accurate. As the implementation of AI is constantly changing and developing, the stance towards it changes too. There may also be regional differences regarding the attitude toward AI. This paper will partly base its


respondents from twelve different countries, of which 4.41% were from the Netherlands.

Current AI literature focuses primarily on skilled versus unskilled workers in the labor force (Upadhyay et al., 2021). Upadhyay et al. (2021) identified a literature gap regarding the effect of technological trends and the motivation behind the intention to accept AI by entrepreneurs.

This study will further explore this gap, and apply their framework to a more specific region, the ecosystem of Amsterdam. The findings may demonstrate differences due to the location of the sample. In contrast, the findings may confirm the same conclusions and show that their framework is also applicable to other settings. This study will test their generalizability.

Furthermore, the results from this thesis can be used by other academics that are specifically interested in AI developments in Amsterdam. As Amsterdam has a very innovative and dynamic ecosystem, many ongoing investigations exist around it. Once it is clear why and where AI is accepted and is not accepted, other researchers can explore how to reduce resistance or further test the validity of this study and its conclusions.

1.2.2 Practical Relevance

With the ever-changing external environment in which firms must survive and compete, they must be aware of current developments, especially those relating to Artificial Intelligence.

This paper will give insight into the stance on AI of start-up owners. The start-up owners are the entrepreneurs of the organization. In addition, entrepreneurship is considered a major driver behind economic and technological changes (Murphy et al., 2006). Start-up companies contribute greatly to this entrepreneurial drive.

By reading this paper, organizations in Amsterdam can get more informed regarding the technological trends in their environment. It can help them improve their marketing strategy and competitor strategy. The organizations may also adjust their stance on the implementation of AI, knowing the viewpoint that the majority of their competitors have. Additionally, the findings from the thesis can be utilized by start-ups waiting to enter the market and give them a clearer understanding of the already existing environment.

Moreover, this study is not only helpful to organizations but also to employees.

Organizations will not use their employees for the tasks their technology can already do. If employees have a greater awareness of the motivational forces behind the intention to accept AI, they can also better accept the technological and disruptive developments in the labor force.

Employees should learn to work with AI rather than against it.


1.3 Research question

By conducting surveys of start-up owners in Amsterdam about their trust in AI, the following research question will be answered:

“How can the intention to accept Artificial Intelligence of Start-up Owners in Ecosystem Amsterdam be explained?”

In relation to this research question, a couple of sub-questions have been formulated to measure the intention to acceptance of Artificial Intelligence (AI):

1. How is performance expectancy related to the intention of start-up owners to accept AI?

2. How is openness related to the intention of start-up owners to accept AI?

3. How is inconvenience related to the intention of start-up owners to accept AI?

4. How is uncertainty related to the intention of start-up owners to accept AI?

5. How is attitude related to the intention of start-up owners to accept AI?


The literature review gives an overview of the key concepts used in this thesis, and the proposed hypotheses. The main argument described here relates to the motivations behind the intention to accept AI.

2.1 Artificial Intelligence

Artificial intelligence (AI) is the technology that can make machines do tasks that require some sort of human intelligence (Neelam, 2022). McCarthy (2007) defines Artificial Intelligence (AI) as "the science and engineering of making intelligent machines." After WWII, AI research began and became exponentially more popular among researchers (McCarthy, 2007). Nowadays, it is already widely used in many fields of work, such as manufacturing, healthcare, and customer services. The implementation of AI makes it possible to use various key cognitive technologies such as machine learning algorithms, robotic process automation, speech recognition, and natural language processing (Kishnani et al., 2017). AI can be used to solve complex problems, and in some cases, it can exceed human capabilities. Ergo, some


question to what extent AI will endanger certain human rights, such as privacy. Or whether AI will take over jobs and create a problem in the labor market (Bolton et al., 2018).

There are two types of AI – symbolic and neural (Upadhyay et al., 2021). Symbolic AI was invented first; it makes the machine think within specific rules and quickly find solutions.

This type of AI is relatively easy to understand and interpret for its human users. On the contrary, neural AI uses a more complex system known as Deep Neural Networks (DNN), which applies a self-learning algorithm, such as facial recognition (Barredo Arrieta et al., 2020;

Upadhyay et al., 2021). The more practice data it is exposed to, the more it learns. Both are important in practice, but this thesis will primarily focus on neural AI. Neural AI is more novel, so fewer people are familiar with it, and for that reason, the chance of resistance against it is higher.

To establish the motivations for AI acceptance, the participants must have an adequate understanding of AI in the first place. Therefore, the participants should be provided with a clear explanation. Explainable AI ensures the effectiveness of the AI system, allowing users to correctly interpret and trust the AI (Barredo Arrieta et al., 2020). This concept raises the question; how should AI be explained to achieve a pragmatic understanding of AI? The paper

"Metrics for explainable AI: Challenges and prospects" addressed that question (Hoffman et al., 2019). The effectiveness of the explanation depends on the knowledge the reader already has. The explanation should induce curiosity but not leave too many loose ends because that may lead to confusion.

2.2 Start-up Owners

The owners of start-ups, also known as start-up entrepreneurs, can use their creative abilities to exploit entrepreneurial opportunities (Arora et al., 2016; Morris et al., 2010). They provide the vision, ensure the organization's product or service production is creative and innovative, and combine that with patience and managerial skills (Morris et al., 2010).

Furthermore, they allocate their resources to capitalize upon a window of opportunity. Thus, their opinion can be considered the most relevant opinion within the organization. Surely, they operate with a team, but they make the most crucial decisions. As AI technology further develops, companies will grow and will have access to more technology.

Entrepreneurial personality is still widely researched, but many studies have had inconclusive results (Mitchell et al., 2002). Certain differences result in some people recognizing an opportunity and others not (Baron, 1998). However, no typical personality traits


can predict a successful entrepreneur. Entrepreneurs view business situations as having more strengths, opportunities, and potential gains than non-entrepreneurs (Palich & Ray Bagby, 1995). Still, the conclusion drawn from various literature is that entrepreneurs can be of all shapes, sizes, and ethnicities (Hatten & Coulter, 1997). A more recent study by Zhao, Seibert and Lumpkin (2010), argued that personality does play a role in the emergence and success of entrepreneurs. They have gathered evidence that four of the Big Five personality traits are significantly related to entrepreneurial intentions and performance. Openness and conscientiousness are the most significant, while agreeableness is mostly unrelated. Individuals that score lower on openness and conscientiousness may find an entrepreneurial career less satisfying.

The level of entrepreneurship can be defined as innovativeness, risk-taking, and proactiveness (Morris et al., 2010). The degree of innovativeness of the entrepreneur depends on individual factors and the environment (Koellinger, 2008). Firstly, education, unemployment, and self-confidence are positively related to entrepreneurial innovativeness on the individual level. Secondly, imitation-type innovations are more likely in developing countries on the environmental level. At the same time, more disruptive innovations are more likely in more economically developed countries that also score higher on the worldwide production possibility frontier (PPF). When an entrepreneur engages in more innovative entrepreneurial activities, he is more likely to accept and use novel technology. The level of AI acceptance in start-up owners is predicted to be relatively high, but its motivations are not foreseeable.

2.3 Start-ups in Ecosystem Amsterdam

This thesis will study the acceptance intention of start-up owners in the entrepreneurial ecosystem of Amsterdam. This ecosystem has a high density of start-ups, and there are constantly new developments and trends to study. New developments bring changes and challenges. This paper will research the AI challenges that come with these developments in ecosystem Amsterdam. Currently, there are roughly 3000 start-ups in Amsterdam, accumulating to an ecosystem value of 223,3 billion euros, which may be partly due to the support from investors (Dealroom.Co, 2020). Amsterdam is also third in Europe in the tech ecosystem value. That is why start-ups in Amsterdam's entrepreneurial ecosystem can be considered a relevant subject to study developments regarding AI.


Moreover, what is an entrepreneurial ecosystem? Firstly, it consists of the term entrepreneurial, which means recognizing opportunities for creating new goods and services (Shane & Venkataraman, 2000). Secondly, the term ecosystem comes from biology, defined as “a biotic community, its physical environment, and all the interactions possible in the complex of living and nonliving components” (Tansley, 1935). An essential part of this definition is the interaction with its environment. The entrepreneurs in the entrepreneurial ecosystem are dependent on the resources and institutions in that environment, as are those resources and institutions influenced by the entrepreneurs (Stam & van de Ven, 2021). In conclusion, an entrepreneurial ecosystem consists of elements that sustain entrepreneurship in a particular area.

Figure 1: Map of the delineated ecosystem in the Netherlands (Stam et al., 2016)

Ecosystem Amsterdam, also known as the Amsterdam Metropolitan Area (MRA), which consists of IJmond, Zaan region, Agglomeration Haarlem, Great-Amsterdam, The Gooi and Vecht region and Flevoland (Stam et al., 2016). It is one of the five entrepreneurial ecosystems in the Netherlands. It is in a very urbanized and highly economically developed area known as the "Randstad".

Start-ups are likely to participate in the technological trends as their innovation rates are generally very high (Hashai & Markovich, 2017). For that reason, studying the relationship between AI and start-up organizations is interesting. Start-ups are often identified by their early stage of organizational development. They generally share characteristics such as high risk, and high growth, driven by creativity and innovation. Blank and Dorf (2012) define start-ups in their start-up manual as "a temporary organization in search of scalable, repeatable,


profitable business model." A start-up can also simply be defined as a "company designed to grow fast" (Graham, 2012). Start-ups usually focus on growth, rapid iteration, and the exploitation of technology to feed new demands in the market (Terryn, 2021). According to, start-ups consist of less than 50 people. According to this rule, there are about 3000 start-up organizations in the entrepreneurial ecosystem of Amsterdam.

2.4 Artificial Intelligence Acceptance Intention

To determine the AI acceptance intention (AIAI), it should be clear what its motivations are. The paper "Theorizing Artificial Intelligence and Digital Entrepreneurship Model"

identified ten elements of the intention to accept (Upadhyay et al., 2021). Those elements together make up the AIDAE model; the entrepreneur’s intention to accept AI, which consists of (1) openness, (2) affordances, (3) generativity, (4) social influence, (5) hedonic motivation, (6) effort expectancy, (7) performance expectancy, (8) attitude, (9) uncertainty, and (10) inconvenience. According to the research done by Upadhyay et al. (2021), the most significant antecedents are performance expectancy, openness, inconvenience, uncertainty, and attitude.

Those five elements are selected for this study, which results in a simplified version of the AIADE model.

2.4.1 Performance Expectancy

Performance expectancy (PE) refers to “the perceived benefits that an individual receives through technology” (Upadhyay et al., 2021). Venkatesh et al. (2003) argued that performance expectancy is the most decisive factor in using new technology. He defined performance expectancy as“the degree to which an individual believes that using the system will help him or her attain gains in job performance”. Additionally, according to the research done by Upadhyay et al. (2021), performance expectancy has the highest positive correlation with the AI acceptance intention. Based on these findings, the following hypothesis is proposed:

Hypothesis 1: Performance expectancy is positively related to the intention of start-up owners to accept AI.


2.4.2 Openness

Openness (OP) refers to “the technology’s features and functionality to facilitate the actor’s participation, contribution, process and outcomes” (Upadhyay et al., 2021). AI is fundamental in developing entrepreneurial innovation, and openness to new technology facilitates that process. AI-driven ecosystems contribute to a more efficient organization; it also benefits individuals and society at large. Innovative ecosystems, such as MRA, can help with collaboration, governance, and decision-making (Upadhyay et al., 2021; Wareham et al., 2013). Upadhyay found that openness was highly correlated to the entrepreneur's intention to accept AI. Based on this information, the following hypothesis is proposed.

Hypothesis 2: Openness is positively related to the intention of start-up owners to accept AI.

2.4.3 Inconvenience

Inconvenience (IN) is the perception of the user on whether the “use of a system or technology is inconvenient due to its hidden features or functions” (Upadhyay et al., 2021).

When users experience inconvenience with the usage of AI when looking for potential solutions, it negatively affects their intention to keep using AI. Upadhyay et al. (2021) found that inconvenience has a high negative correlation to the entrepreneur's intention to accept AI.

Based on these findings, the following hypothesis is proposed.

Hypothesis 3: Inconvenience is negatively related to the intention of start-up owners to accept AI.

2.4.4 Uncertainty

Uncertainty (UN) refers to the “individuals’ uncertain perception of technology or system in context related to the actual outcome” (Upadhyay et al., 2021). Market uncertainty can affect entrepreneurial initiatives such as implementing new technology. AI is a disruptive innovation that can greatly influence the job and workforce (Girasa, 2020). Especially neural AI is very advanced and unfamiliar; therefore, there is a perceived uncertainty with the widespread acceptance of neural AI. Individuals that perceive a high level of market uncertainty, may be less likely to use the new technology. This leads to the following hypothesis:


Hypothesis 4: Uncertainty is negatively related to the intention of start-up owners to accept AI.

2.4.5 Attitude

Attitude (AT) refers to “an individual’s liking or feeling of a particular behavior”. In the context of this paper, it concerns the attitude toward AI. Upadhyay et al. (2021) found that attitude is positively related to the intention to accept AI. Based on these findings, the following hypothesis is formed:

Hypothesis 5: The attitude towards AI is positively related to the intention of start-up owners to accept AI.


Figure 2: Conceptual Model



The methodology section will describe the methods used to research the AI acceptance intention of start-up owners. Firstly, the empirical setting and data collection method will be presented. Consequently, the independent, dependent, and control variables will be defined.

Lastly, the analytical procedure is described.

4.1 Empirical setting and data collection

This study will follow a quantitative method in which start-up owners fill in surveys.

There are about 3000 start-ups in the ecosystem of Amsterdam. The goal is to get a sample of at least 60 participants but striving to exceed that number. Purposive sampling was used to find the respondents, so only start-up owners in ecosystem Amsterdam were approached. A start- up is defined as an organization with less than 50 people. Additionally, the start-up entrepreneurs own a business that is active within the ecosystem of Amsterdam. During the gathering of responses, some remote workers were contacted; if they owned a business that was not active in Amsterdam, but they do live in Amsterdam, they do not fit the requirements.

Furthermore, to obtain the desired number of responses, LinkedIn Sales Navigator and SalesFlow were used. SalesFlow can be used to set up campaigns to send automated messages via LinkedIn to the people selected through Sales Navigator. Sales Navigator allows one to perform a bullion search to filter businesses and employees based on location, job title, years in current position, size of the company, main operations, and keywords. For this bullion search, location filters were added by selecting the main cities within the ecosystem of Amsterdam. For job titles, “owner” and “founder” had been selected. Lastly, for years at the company, 1-2 years were used.

Moreover, about 2500 LinkedIn members were found with this search, of which 344 were invited over a period of 3 weeks, which means they got an automated introduction message along with a connection request. However, even with this bullion search, there will be a margin of error. Some will not fit the requirements, which is why in the follow-up message the requirements will also be explained, and again in the survey. Of those 344, 208 accepted the request, which means they got a follow-up message explaining the research topic and stating the requirements. Finally, 130 people replied to the message, of which 87 filled in the survey. However, only 69 participants finished the entire survey, resulting in a useable sample of 69.


4.2 Variables and measures

This study uses the ADIAE framework by Upadhyay et al. (2021), from which the five most significant factors related to AI acceptance are selected. The intention to accept AI (AIAI) is the independent variable of this quantitative study, and the five antecedents, (1) performance expectancy, (2) openness, (3) inconvenience, (4) uncertainty, and (5) attitude, are the dependent variables. A survey was created to generate the data for the analysis. Like the survey used for the original ADIAE model, three questions had to be answered for openness, inconvenience, and uncertainty, and four questions for performance expectancy, attitude, and AI acceptance intention. All participants were start-up entrepreneurs that own a business located within the entrepreneurial ecosystem of Amsterdam. Correspondingly to the ADIAE model, each question was answered based on a 7-point Likert scale; (1) strongly disagree, (2) disagree, (3) more or less disagree, (4) neutral, (5) more or less agree, (6) agree, (7) strongly agree. All questions were regarding the personal perception of the respondent of each variable. This study will test how well the independent variables can predict the outcome of the dependent variable, AI acceptance intention. Additionally, at the start of the survey, a brief explanation of neural AI was given. The respondents had to answer whether they understood since the questions cannot be reliably answered if they do not understand the topic.

4.3 Control Variables

Furthermore, the respondents also had to answer some questions related to the demographics, such as age, education, nationality, and the job sector they were active in. As seen below in figure 3, figure 4, most participants completed a master’s degree, and most were aged above 40 (see Appendix A1). 81.2 percent of the respondents had a Dutch nationality, while the others were expats. The most popular job sectors were (1) business, consulting, and management, (2) information technology, (3) recruitment and HR, and (4) marketing, advertising, and PR.


4.4 Analytical procedure

Finally, for this quantitative research design, SPSS was used to analyze the data. This statistical data program was also used to screen, recode and clean the data. Descriptive statistics provide some general information about the variables, such as the sample size, mean, and standard deviation. Lastly, linear regressions were used to test the five hypotheses.


For the results section, SPSS is used to perform the analysis. The goal is to test the effect of the predictor variables (performance expectancy, openness, uncertainty, inconvenience, and attitude) on the outcome variable (AI acceptance intention) based on a sample of 69. The hypotheses are tested via a multivariate linear regression analysis. Moreover, in order to perform a regression analysis, various preparations must be made. Firstly, the data must be screened to check the quality of the data and make sure that there are no impossible answers. Then, a reliability analysis is performed to check the internal consistency of the items per variable. If the reliability analysis is done, the variables need to be computed to use for the rest of the analysis.

Furthermore, some linear regression assumptions must be tested before doing the actual analysis. The linear relationship, multivariate normality, and multicollinearity assumptions are tested. Finally, the regression can be performed, and the hypotheses are tested. Additionally, some alternative regressions will be performed with different control groups.

5.1 Data cleaning, recoding, and correlations

The data is screened and cleaned, meaning that samples with incomplete or impossible answers were deleted as well as samples that did not fit the requirements, which resulted in a total sample of 69. In the questionnaire, there were 2 reverse-items which were recoded. Item 26, “I think that AI technology will be easy to use and implement,” corresponds to the independent variable, inconvenience. Furthermore, Item 37, “I feel that I have resistance against the AI technology,” corresponds to the dependent variable, AI acceptance intention.

5.1.1 Reliability analysis

Consequently, a reliability analysis was performed to test the internal consistency. To test the reliability of the variables and their corresponding items (with n=69), Cronbach’s Alpha


() is calculated with SPSS. Cronbach’s Alpha depends on the number of items on the scale (Field, 2017). The top half of the equation contains the number of scale items squared;

therefore, if that number of questions is greater, Cronbach’s Alpha will also be greater even if the questions have a lower average correlation. This study uses a 7-point Likert scale for all the questions. The value should be between 0.7-0.8 for good internal consistency and 0.6 as the threshold value (Field, 2017). Potentially, the value of Cronbach’s Alpha can be increased if certain items are deleted.

Performance expectancy generated a Cronbach’s Alpha (4 items; =.666) that surpasses the minimum value of 0.6 but is not greater than 0.7, which is needed for a good internal consistency reliability. Cronbach’s Alpha cannot be increased with the deletion of an item. Additionally, openness (3 items; =0.655) and attitude (4 items; =.674) generated a Cronbach’s Alpha between 0.6 and 0.7. Similarly to performance expectancy, openness’

Cronbach’s Alpha cannot be increased by deleting any items. While Cronbach’s Alpha of Attitude can be increased to .685 with the deletion of an item, this improvement is so small that it is not worth deleting the item.

Furthermore, the reliability analysis for inconvenience resulted in a Cronbach’s Alpha (3 items; =.553) below 0.6, which is below the minimum value. Cronbach’s Alpha can be increased with the deletion of one item to .621, but theoretically speaking, that would not make sense because only two items would be left. Likewise, uncertainty also generated a Cronbach’s Alpha (3 items; =.508) below 0.6. Again, Cronbach’s Alpha can be improved by deleting an item, leaving only two items total for this variable. No items were deleted. Lastly, AIAI generated a good Cronbach’s Alpha (4 items; =.732), which indicates high internal consistency.

To conclude, the results should be interpreted with caution because not all variables have high internal consistency. However, these Cronbach’s Alpha values may also be due to the low number of questions per variable. Finally, all the corresponding items are merged together to create the six different variables. The variables are created by taking the average value of the items.

5.2 Descriptives and regression assumptions

Some assumptions that need to be checked to perform the regression analysis (Field, 2017). Most independent variables are significantly correlated, which may imply


outliers which may make the regression analysis less reliable. Before continuing with the regression, the linear relationship, multivariate normality, and multicollinearity are tested to ensure that the conclusions from the data are accurate.

Table 1. Descriptive statistics and correlations

M SD 1 2 3 4 5

1. PerformanceExpectancy (PE) 5.74 .79 (.666)

2. Openness (OP) 5.47 .99 .467** (.655)

3. Inconvenience (IN) 3.77 1.15 -.308* -.518** (.553)

4. Uncertainty (UN) 4.03 1.21 -.414** -.273* .294* (.508)

5. Attitude (AT) 6.31 .60 .517** .337** -.055 -.125 (.674)

** Correlation is significant at the 0.01 level (2-tailed).

* Correlation is significant at the 0.05 level (2-tailed).

(n = 69)

Chronbach’s alphas on diagonal 5.2.1 Linear relationship

There should be a linear relationship between the independent and dependent variables.

When looking at the linearity of the relationship between the independent and dependent variables, it can be best visually screened with a scatter plot (Field, 2017) (see Appendix A2).

A clear linear relationship is visible between AI acceptance intention (DV) and performance expectancy (IV). The value of R Square is 0.497, which indicates that performance expectancy can explain 49.7% of the variation in AIAI. Also, a linear relationship is portrayed between AI acceptance intention (DV) and openness (IV). The value of R Square is 0.323, which suggests that openness can explain 32.3% of the variation in AIAI.

Furthermore, there is a weak negative linear relationship between AI acceptance intention (DV) and uncertainty (IV). The value of R Square is 0.073, which indicates that uncertainty can explain only 7.3% of the variation in AI acceptance intention.

Additionally, there is a positive linear relationship between AI acceptance intention (DV) and attitude (IV). The value of attitude’s R Square is 0.232, which indicates that attitude can explain 23.2% of the variation in AIAI. Moreover, a negative linear relationship between AI acceptance intention (DV) and inconvenience (IV) can be seen in the scatterplot. The value of R Square is 0.122, which suggests that inconvenience can explain 12.2 % of the variation in AIAI.

All the scatterplots suggest a linear relationship, but the Z-values can be used to look more in-depth at whether there are problematic outliers. Outliers are values that are drastically different from the rest of the scores (Field, 2017). Those outliers may drive significant findings,


which evidently result in unreliable conclusions. The Z-values are calculated by SPSS, they show scores in terms of a distribution with a mean of 0 and standard deviation of 1. There was only one extreme case (z = 3.84) within attitude, which can therefore indicate an outlier. In table 2, the outliers of attitude are computed in a table to investigate this potential issue more in-depth. According to the book about SPSS by Field (2017), about 5% or less are to be expected to be potential outliers (z >1.96), less than 1% should be probable outliers (z >2.58), and very few cases above 3.29. The remaining cases (95.8%) are within the normal range.

These percentages are mostly consistent with a normal distribution; therefore, the outlier was not excluded from the analysis.

Table 2. Zscores - Attitude


Valid Percent

Cumulative Percent

Valid Extreme (z-score > 3.29) 1 1.4 1.4

Probable outliers (z > 2.58) 1 1.4 2.9

Potential outliers (z > 1.96) 1 1.4 4.3

Normal range 66 95.7 100.0

Total 69 100.0

5.2.2 Multivariate normality

When performing the multivariate linear regression, it is assumed that the residuals are multivariate normally distributed (Field, 2017). Skewness and kurtosis should be between -1 and +1 for multivariate normality. Skew refers to the lack of symmetry in the distribution, and kurtosis refers pointiness of the distribution. Performance expectancy, openness, uncertainty, and inconvenience have skewness and kurtosis within acceptable ranges (see table 3). Attitude and AI acceptance intention have a positive kurtosis, meaning it has a leptokurtic or heavy- tailed distribution.

Table 3. Skewness and kurtosis (n = 69)


Skewness -.299 -.699 .020 -.961 -.866 -.303

Std. Error of Skewness .289 .289 .289 .289 .289 .289

Kurtosis -.568 .733 -.245 1.877 2.222 .262

Std. Error of Kurtosis .570 .570 .570 .570 .570 .570


5.2.3 Multicollinearity

There should be little to no multicollinearity within the independent variables for a dependable regression. When independent variables are highly correlated, multicollinearity can occur. This high correlation between predictors can make it hard to determine the individual significance of the predictors. To check the multicollinearity of the variables, the correlation coefficients, the variance inflation factor, and the tolerance are inspected.

All the correlation coefficients of the independent variables need to be substantially smaller than 1; fortunately, this is the case as seen in table 1. Moreover, if tolerance (T) is below 0.1 and the variance inflation factor (VIF) is greater than 5, there may be an indication of multicollinearity within the data (Field, 2017). Tolerance (T=1-Rsquare) is 0.415 (T=1- 0.585), which does not indicate multicollinearity among the independent variables. Variance Inflation Factor (VIF= 1 / T) is 2.41 (VIF= 1 / 0.451) in this case, which also does not indicate multicollinearity. To conclude, there is no clear indication of multicollinearity in the data.

Therefore, the linear regression can be performed to test the hypotheses.

5.2.4 Likert scale

There is an ongoing conversation about the Likert scale and whether it can be considered ordinal or interval. An ordinal scale cannot be used for statistical analyses.

Officially, a Likert scale is an ordinal scale. In the past, researchers have stated that treating an ordinal scale as an interval scale to perform a statistical analysis can still be very insightful (Knapp, 1990; Stevens, 1946). Therefore, the data from this paper will be treated as an interval scale to continue with the statistical analysis.

5.3 Linear regression

A multiple regression was conducted to test the hypotheses (see Appendix A3). The linear regression will generate the data to check the significance of the individual predictors as well as the significance of the overall model. The regression was repeated separately for the largest age group (see Appendix A4) and education group (see Appendix A5) to see whether there are any significant differences in findings with these control groups.

The overall model with the five predictors explains 58.5% (Rsquare = .585) of the variation in AI acceptance intention, with a significance of p<.001 (see Appendix A3). The adjusted R square can indicate how well the model generalizes (Field, 2017), which is 55.2%

(adjusted Rsquare = .552) in this case.


5.3.1 Hypotheses

Table 4 shows that performance expectancy has the highest impact on AI acceptance intention (H1: PE→AIAI; p<0.001, B=.615). Therefore, there is enough statistical evidence to support hypothesis 1. Performance expectancy is positively related to AI acceptance intention.

Additionally, the openness construct is also significant (H2: OP→AIAI; p<0.05, B=.253).

There is enough statistical evidence to support hypothesis 2 as well, which means that openness is positively related to AI acceptance intention. However, the relationships between the constructs of uncertainty, inconvenience, and attitude with AI acceptance intention have inconclusive results. While the mean of Attitude is relatively the highest (M=6.31), it’s effect on AI acceptance intention is insignificant. There is insufficient statistical evidence to reject the null hypotheses (p>0.05); therefore, hypothesis 3, hypothesis 4, and hypothesis 5 are not supported. To conclude, performance expectancy and openness are the best constructs for predicting AI acceptance intention. If respondents expect the performance of AI technology to be good and if they view the AI technology to be open and accessible to them, they are more likely to accept the new AI technology.

5.3.1 Alternative regressions

As described in the methods section, there were visibly a majority of some demographic groups (see Appendix A1). Most respondents were over the age of 40, and most had obtained a master’s degree (see Figure 3 and Figure 4). It may be interesting to see whether there are any significant changes in outcomes with those control groups. It may or may not indicate

Table 4. Coefficients a


Unstandardized Coefficients





Error Beta t Sig.

1 (Constant) -.510 1.043 -.490 .626

PE .615 .131 .517 4.713 .000

OP .253 .100 .266 2.539 .014

UN .038 .071 .049 .540 .591

AT .199 .153 .128 1.298 .199

IN -.049 .080 -.060 -.607 .546

a. Dependent Variable:AIAI


When looking at the results with respondents over 40, the explained variation increases to 62.4% (Rsquare= .624) (see Appendix A4). There are no significant changes in the individual significance of the predictors. Moreover, when looking at the results with respondents that have a completed education of a master’s and/or doctorate degree, the explained variation of the model increases to 73.8% (Rsquare= .738) (see Appendix A5). There are no significant changes in the individual significance of the predictors. Likewise, the adjusted Rsquare of both alternative regression also increase, indicating better generalizability of the model with these control variables.


The research question is as follows: “How can the intention to accept Artificial Intelligence of Start-up Owners in Ecosystem Amsterdam be explained?”. According to the results, openness and performance expectancy are significant predictors of AI acceptance intention, two of the five hypotheses are supported. Performance expectancy explains 49.7%

of the variation, and openness explains 32.3%. The data indicate that performance expectancy is the most important predictor. The overall model explains 58.5% of the variation in the outcome variable, AI acceptance intention. The significance of the model is increased when respondents have attained at least a master’s degree; the explained variation increases from 58.5% to 73.8%. Also, the explained variation of the model increases to 62.4% with respondents over age 40. When looking at the regression and using a control variable, the model becomes better at predicting. When using the responses of participants over the age of 40 and participants that have obtained at least a master’s degree, the adjusted Rsquare is also increased.

These findings indicate that education and age may play a mediating role in the model, but further research should be conducted for more conclusive findings.

Similarly to the study done by Upadhyay et al. (2021), openness and performance expectancy are the most significant predictors. However, attitude, inconvenience, and uncertainty were also significant in their study. This difference in findings might be mostly explained by the questions in the survey. While the style and format of the questionnaire were similar, different questions were used. Therefore, this thesis is not an exact replica; it merely uses an adapted ADIAE model. Additionally, in this study the demographics were different.

As planned, the empirical setting of this study was set within the entrepreneurial ecosystem of Amsterdam. While in the Upadhyay et al. (2021) study, respondents were from many different countries, of which only 4% were from the Netherlands. The participants of this study were


mostly over age 40, while in the study by Upadhyay et al. most of the participants were age 31-40. Additionally, most respondents from this study had obtained at least a master’s degree (47.8%), while most respondents in the Upadhyay et al. study had only obtained a bachelor’s degree (65.33%). Cultural differences may explain these particular outcomes.

6.1 Limitations and further research

This study has, like any other study, its own limitations. When performing the reliability analysis, Cronbach’s alpha was not for every variable above the threshold value. At the same time, items were not deleted even if they could increase Cronbach’s alpha slightly because there are only 3 to 4 items per variable. Cronbach’s alpha measures the internal consistency of those items; it tests whether all the items measure the same concept (Tavakol & Dennick, 2011). In this case, a low Cronbach’s alpha may be due to the low number of questions.

The model generated an outlier, which was not deleted because over 95% was still within the normal range. Additionally, AIAI and attitude have a high positive kurtosis (kurtosis

> 1). A recommendation for future research is to increase the sample size, to make the distribution more normal (Field, 2017). Given the limited time and resources, a sample size of only 69 could be used for the analysis. If researchers want to replicate this study with a bigger sample size, taking more time to gather all the responses is essential. Only 25% of the people contacted were willing to participate. Also, keeping in mind that not all of those respondents will finish the survey, since about 20% of the respondents of this study did not finish the entire survey. Consequently, those responses have been excluded from the analysis.

Future researchers can also conduct qualitative analysis to further test the theorized model. Qualitative research can be used to create a more in-depth understanding of different cultural factors. Additionally, it may be interesting to pay more attention to the different age groups and educational attainment levels, as these demographic groups influenced the explained variation of the overall model. Moreover, it may also be meaningful to use the same research model and focus on different industries. This study did not obtain enough responses to draw reliable conclusions about the influence of the type of industry on the AI acceptance intention. Lastly, if future researchers want to replicate this quantitative study, adding more questions per variable may be wise to confirm good internal consistency.


6.2 Implications

This paper investigates the motivations of start-up owners in ecosystem Amsterdam to accept AI. It contributes to the academic progression of AI, particularly relating to entrepreneurship. By developing a simplified theoretical model based on the ADIAE model by Upadhyay et al. (2021), technology acceptance theories are expanded. This study highlights the importance of a start-up entrepreneur’s willingness to accept AI for digital entrepreneurship. Additionally, it allows conducting further research into the antecedents of AI acceptance intention. Furthermore, this paper generated findings on start-up entrepreneurs’

intention to accept AI based on empirical data from 69 participants.

The research findings also have implications for practitioners. Policymakers, marketeers, and public agencies can benefit from understanding the entrepreneurs’ intention to accept AI. They should acknowledge the importance of openness and performance expectancy.

Considering that AI implementation can provide great national benefits, many countries have started strategic plans for effective digital entrepreneurship (Upadhyay et al., 2021).

Government agencies in the Netherlands that wish to promote AI acceptance within the ecosystem of Amsterdam should incentivize the entrepreneurs that are open to the usage of AI to promote their experiences publicly. When other entrepreneurs are exposed to more AI experiences, they are more likely to become more open to new technology, and their perceived performance expectancy of AI may be increased, facilitating the AI adoption. Additionally, policymakers should examine the attractiveness of AI ecosystem services, solutions, products and knowledge sharing for entrepreneurs. The government can also provide monetary incentives for AI-entrepreneurship R&D activities and human training for better AI acceptance within the ecosystem.

Furthermore, businesses in Amsterdam Metropolitan Area may also find the results from this thesis valuable. Businesses can use the information to get a better understanding of their competitors, and why some businesses are more open to new technology. The intention to accept AI is mostly determined by the perceived openness of the technology and performance expectancy. If there are multiple leaders in a business, there may be conflicts regarding a new strategy relating to the technology. The findings from this study can be used to understand the different partners and can perhaps indicate the source of the willingness to accept or reject new AI technology. Moreover, employees may find this research paper beneficial as well. Reading this research may show the technological trends in start-up businesses. Therefore, illustrating that technical skills and knowledge regarding AI are


becoming more important and valuable. Having skills and knowledge about AI technology may provide employees with more business opportunities and secure their position in the future workforce.

6.3 Conclusion

The undertaken study found significant results for openness and performance expectancy as predictors for AI acceptance intention. Uncertainty, inconvenience and attitude did not show any significant relationship. However, further research should be conducted for more generalizability, particularly with a greater sample. It is valuable to emphasize the importance of AI technology adoption since it can provide great national benefits.

Governments may use the data to facilitate better AI adoption. Also, businesses should keep track of these technological developments to stay active in the competitive landscape. Even employees can use these findings to their advantage and ensure a better fit in the future labor market. Additionally, this research contributes to current AI literature and literature regarding the technological trends within the ecosystem of Amsterdam.


Arora, A., Cohen, W. M., & Walsh, J. P. (2016). The acquisition and commercialization of invention in American manufacturing: Incidence and impact. Research Policy, 45(6), 1113–1128.

Baron, R. A. (1998). Cognitive mechanisms in entrepreneurship: Why and when

enterpreneurs think differently than other people. Journal of Business Venturing, 13(4), 275–294.

Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., Chatila, R., & Herrera, F.

(2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies,

opportunities and challenges toward responsible AI. Information Fusion, 58, 82–115.


Bolton, C., Machová, V., Kováčová, M., & Valášková, K. (2018). THE POWER OF HUMAN–MACHINE COLLABORATION: ARTIFICIAL INTELLIGENCE, BUSINESS AUTOMATION, AND THE SMART ECONOMY. Economics, Management, and Financial Markets, 13(4), 51–56.

Business Insider, N. (2020, December 16). In deze 3 sectoren levert kunstmatige intelligentie voorlopig de meeste waarde op. Business Insider Nederland. nederland/ (2020, September 8). Dealroom.Co. europes-3-tech-ecosystem

Field, A. (2017). Discovering statistics using IBM SPSS Statistics. (5th ed.). Sage.

Girasa, R. (2020). AI as a Disruptive Technology. In R. Girasa (Ed.), Artificial Intelligence as a Disruptive Technology: Economic Transformation and Government Regulation (pp. 3–21). Springer International Publishing. 35975-1_1

Graham, P. (2012). Startup = Growth.

Hashai, N., & Markovich, S. (2017). Market Entry by High Technology Startups: The Effect of Competition Level and Startup Innovativeness. Strategy Science, 2(3), 141–160.

Hatten, T. S., & Coulter, M. K. (1997). Small Business: Entrepreneurship and Beyond.

Prentice Hall.

Hoffman, R. R., Mueller, S. T., Klein, G., & Litman, J. (2019). Metrics for Explainable AI:

Challenges and Prospects. ArXiv:1812.04608 [Cs]. Kishnani, P., Eggers, W. D., & Fishman, T. (2017). AI-augmented human services: Using

cognitive technologies to transform program delivery (p. 24). Deloitte Center for


Government Insights. services/4152_AI-human-services.pdf

Knapp, T. R. (1990). Treating Ordinal Scales as Interval Scales: An Attempt To Resolve the Controversy. Nursing Research, 39, 121–123.

Koellinger, P. (2008). Why are some entrepreneurs more innovative than others? Small Business Economics, 31(1), 21. McCarthy, J. (2007). WHAT IS ARTIFICIAL INTELLIGENCE? Computer Science

Department. Stanford University., 15.

Mitchell, R. K., Busenitz, L., Lant, T., McDougall, P. P., Morse, E. A., & Smith, J. B. (2002).

Toward a Theory of Entrepreneurial Cognition: Rethinking the People Side of Entrepreneurship Research. Entrepreneurship Theory and Practice, 27(2), 93–104.

Morris, M. H., Kuratko, D. F., & Covin, J. G. (2010). Corporate Entrepreneurship &

Innovation (3rd ed.). Cengage Learning Publishing.

Murphy, P. J., Liao, J., & Welsch, H. P. (2006). A conceptual history of entrepreneurial thought. Journal of Management History, 12(1), 12–35.

Neelam, M. (2022). Aspects of Artificial Intelligence In Karthikeyan.J, Su-Hie Ting and Yu- Jin Ng (eds), “Learning Outcomes of Classroom Research” p:250-256, L’ Ordine Nuovo Publication, India. 978-93-92995-15-6.

Palich, L. E., & Ray Bagby, D. (1995). Using cognitive theory to explain entrepreneurial risk-taking: Challenging conventional wisdom. Journal of Business Venturing, 10(6), 425–438.


Stam, E., Romme, A. G. L., Roso, M., van den Toren, J. P., & van der Starre, B. T. (2016).

Knowledge triangles in the Netherlands: An entrepreneurial ecosystem approach.


Stam, E., & van de Ven, A. (2021). Entrepreneurial ecosystem elements. Small Business Economics, 56(2), 809–832.

Stevens, S. S. (1946). On the Theory of Scales of Measurement. Science, 103(2684), 677–


Tansley, A. G. (1935). The use and abuse of vegetational concepts and terms. Ecology 16, 284—307. Progress in Physical Geography: Earth and Environment, 31(5), 284—


Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach’s alpha. International Journal of Medical Education, 2, 53–55. Terryn, L. G. (2021, November 2). Dutch startups have created 130,000 jobs in the

Netherlands. Dealroom.Co. Upadhyay, N., Upadhyay, S., & Dwivedi, Y. K. (2021a). Theorizing artificial intelligence

acceptance and digital entrepreneurship model. International Journal of Entrepreneurial Behavior & Research, ahead-of-print(ahead-of-print).

Upadhyay, N., Upadhyay, S., & Dwivedi, Y. K. (2021b). Theorizing artificial intelligence acceptance and digital entrepreneurship model. International Journal of

Entrepreneurial Behavior & Research, ahead-of-print(ahead-of-print).

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly: Management Information Systems, 27(3), 425–478.


Wareham, J., Fox, Ph. D., Paul, & Cano Giner, J. L. (2013). Technology Ecosystem Governance. SSRN Electronic Journal, 25. Wilman, D. (2022, January 7). Artificial intelligence inzetten om werknemers te ontlasten bij

saaie en repetitieve taken—Zo doet het Amsterdamse BrainCreators dat. Business Insider Nederland. taken/

Zhao, H., Seibert, S. E., & Lumpkin, G. T. (2010). The Relationship of Personality to Entrepreneurial Intentions and Performance: A Meta-Analytic Review. Journal of Management, 36(2), 381–404.


APPENDIX A – Figures and tables

Appendix A1 Demographic characteristics Table A1.1 Demographic characteristics (n = 69)

Measure Item Frequency Percentage

Age 25-30 9 13.0

31-35 18 26.1

36-40 12 17.4

above 40 30 43.5

Highest completed education High school 8 11.6

Post secondary vocational education (MBO) 2 2.9

Bachelor's degree (HBO/WO) 21 30.4

Master's degree 33 47.8

Doctorate degree 5 7.2

Marriage status Single 25 36.2

Married 34 49.3

Rather not say 10 14.5

Nationality Dutch 56 81.2

Italian 3 4.3

Polish 2 2.9

American 1 1.4

Brazilian 1 1.4

Colombian 1 1.4

Finnish 1 1.4

French 1 1.4

German 1 1.4

Taiwanese 1 1.4

Ukrainian 1 1.4

Job sector Business, consulting, management 18 26.1

Information technology 11 15.9

Recruitment and HR 10 14.5

Marketing, advertising, PR 7 10.1

Healthcare 4 5.8

Retail 4 5.8

Media and internet 3 4.3

Sales 3 4.3

Hospitality and events 2 2.9

Leisure, sport, tourism 2 2.9

Accountancy 1 1.4

Creative arts and design 1 1.4

Energy and utilities 1 1.4

Science and pharmaceuticals 1 1.4

Social care 1 1.4


Appendix A2 Scatterplots

Figure A2.1 Scatterplot of AIAI and PE Figure A2.2 Scatterplot of AIAI and OP

Figure A2.3 Scatterplot of AIAI and UN Figure A2.4 Scatterplot of AIAI and AT

Figure A2.5 Scatterplot of AIAI and IN


Appendix A3 Regression

Table A3.1 Model summary

Change Statistics

Model R

R Square

Adjusted R Square

Std. Error of the Estimate

R Square Change


Change df1 df2

Sig. F Change

1 .765a .585 .552 .63069 .585 17.727 5 63 .000

a. Predictors: (Constant), IN, AT, UN, OP, PE Table A3.2 ANOVA a


Sum of

Squares df


Square F Sig.

1 Regression 35.257 5 7.051


7 .000b

Residual 25.060 63 .398

Total 60.317 68

a Dependent Variable: AIAI

b Predictors: (Constant), IN, AT, UN, OP, PE Table A3.3 Coefficients a


Appendix A4 Regression – Age > 40

Table A4.1 Model summary – age = above 40 Model

R age>=

31-35 R Square

Adjusted R Square

Std. Error of the Estimate

1 .790a .624 .564 .75478

a. Predictors: (Constant), IN, AT, UN, OP, PE

Table A4.2 ANOVA a,b


Sum of

Squares df


Square F Sig.

1 Regression 22.713 5 4.543 7.974 .000c

Residual 13.673 24 .570

Total 36.385 29

a. Dependent Variable: AIAI

b. Selecting only cases for which age = above 40 c. Predictors: (Constant), IN, AT, UN, OP, PE

Table A4.3 Coefficients a,b


Unstandardized Coefficients



B Std. Error Beta t Sig.

1 (Constant) -2.376 2.012 -1.181 .249

PE .375 .258 .309 1.453 .159

OP .551 .196 .484 2.805 .020

UN .106 .155 .122 .679 .504

AT .425 .279 .251 1.524 .140

IN -.052 .128 -.066 -.403 .691

a. Dependent Variable: AIAI

b. Selecting only cases for which age = above 40


Appendix A5 Regression – Education >= Master’s degree

Table A5.1 Model summary - Education >= Master's degree Mode


R education>= Master's degree

R Square

Adjusted R Square

Std. Error of the Estimate

1 .859a .738 .697 .62802

a. Predictors: (Constant), IN, AT, UN, OP, PE

Table A5.2 ANOVA a,b


Sum of

Squares df


Square F Sig.

1 Regression 35.558 5 7.112 18.031 .000c

Residual 12.621 32 .394

Total 48.179 37

a. Dependent Variable: AIAI

b. Selecting only cases for which education >= Master's degree c. Predictors: (Constant), IN, AT, UN, OP, PE

Table A5.3 Coefficients a,b


Unstandardized Coefficients



B Std. Error Beta t Sig.

1 (Constant) -3.003 1.460 -2.057 .048

PE .727 .163 .539 4.455 .000

OP .327 .138 .327 2.372 .024

UN -.018 .104 -.018 -.171 .865

AT .380 .205 .199 1.854 .073

IN .080 .112 .087 .715 .480

a. Dependent Variable: AIAI

b. Selecting only cases for which education >= Master's degree

APPENDIX B – Survey questions

Block 1 Q1:


This questionnaire uses Qualtrics. The University of Amsterdam (UvA) is committed to protecting your privacy. The University of Amsterdam has a privacy statement, which can be read here (

The objective of this study is to collect information about the motivations behind Artificial Intelligence Acceptance in start-up owners in the entrepreneurial ecosystem of Amsterdam,


also known as Amsterdam Metropolitan Area (MRA).

Requirements to participate

- You are the owner of a start-up organization (<50 people) - The organization is located in the Amsterdam Metropolitan Area Questionnaire

You are being asked to complete a questionnaire that should take approximately 5 minutes to complete.

These data will be used for Lisa's Bachelor Thesis. You may withdraw your consent at any time by contacting Lisa (email or whatsapp +31634771847). Once you have withdrawn your consent, your personal data will no longer be used. Once your personal data are no longer required, they will not be registered by the researcher.


If you have any questions about this study, please address these to Lisa. If you have any questions at a later stage, desire additional information or wish to withdraw your consent, please get in touch with the lead researcher by email or whatsapp. By selecting 'Yes' below, you agree that the above information has been explained to you and that your questions have been answered. You understand that you may ask questions in future about every aspect of this study. By completing and signing this form, you agree to participate in this study.

Do you agree to the details set out in the consent form?

1. Yes 2. No Q2:

Requirements to participate

- You are the owner of a start-up organization (<50 people)

- The organization is located in the Amsterdam Metropolitan Area (see figure above) Do you fit the requirements?

1. Yes 2. No

Q3: What is your age?

1. 18-24 2. 25-30 3. 31-35 4. Above 40

Q4: What is your highest level of completed education?

1. High school

2. Post secondary vocational education (MBO) 3. Bachelors degree (HBO/WO)

4. Masters degree 5. Doctorate degree




Related subjects :