Female Entrepreneurship Facing Artificial Intelligence in the European Union Countries

Artificial Intelligence (AI) is widely regarded as the technology of the future. The intelligent transformation, involving the diffusion of AI technology from developers to enterprises, changes companies' business models and necessitates an understanding of its enablers and obstacles (Wang et al., 2022). Despite the potential benefits of AI, there are also risk factors associated with intelligent transformation, as AI may lead to unemployment, inequality and inhuman outcomes (Moon, 2023). The deepening of socio-economic inequalities is considered one of the threats posed by AI (Wach et al., 2023). These concerns raise a fundamental question: will AI contribute to fostering a more just and equal society, or will it exacerbate various forms of social and economic disparities.

This chapter focuses on the gendered perspective of AI, as there are conflicting arguments about the effect of digitalisation on women's professional development, distinguishing between the digital gender dividend or the digital gender divide (Chen et al., 2021). Specifically, the goal of this chapter is to recognise the relationship between female entrepreneurship and the diffusion of the adoption of AI technologies within business contexts at the macroeconomic level. As the widespread of AI technologies affects all businesses, no matter whether digital or traditional, this study explores the interrelationships between AI and female entrepreneurship, regardless of their activity in the digital or traditional realm.

Understanding how female entrepreneurship responds to AI is crucial for the following reasons. Despite the ongoing progress towards achieving gender equality (as manifested in Sustainable Development Goal 5), entrepreneurship remains one of the domains where a persistent gender gap exists across time and space. Entrepreneurship is a gendered phenomenon, with men significantly more likely to establish and run their businesses than women (Mustafa & Treanor, 2022). In the European Union (EU) countries, women account for approximately 30% of entrepreneurs (Dilli & Westerhuis, 2018; Ughetto et al., 2020).

Furthermore, despite the progress in gender equality, a digital gender divide persists (Lechman & Popowska, 2022; Picatoste et al., 2023; Sánchez-Rivero et al., 2023; Vimalkumar et al., 2021; Yeganehfar et al., 2018). This divide stems from the under-representation of women in the information and communication technology (ICT) sector (in formal roles as ICT specialists and ICT graduates, as in the EU countries, women account for about 20% of employed ICT specialists and students enroled in tertiary education) and the comparatively lower digital proficiency of women relative to men (in informal capacities). The design, creation and implementation of AI technologies require digital competence (Hidalgo et al., 2020).

All these arguments raise concerns about how women entrepreneurs navigate the business environment in the era of AI. Specifically, the research question in this chapter is whether the widespread use of AI becomes another barrier for women entering entrepreneurship due to the gender digital divide. Or rather, since AI technologies are intended to assist humans in their business endeavours, could the widespread adoption of AI serve as an enabling factor for women to venture into entrepreneurship? The scarcity and fragmentation of research on digital female entrepreneurship (Alhajri & Aloud, 2023), together with the significance of addressing gender and digital equality (Picatoste et al., 2023), justifies undertaking the research.

The remainder of this chapter is the following. The discussion on the gendered perspective of entrepreneurship during the digitalisation, including the intelligent transformation, is followed by the presentation of the research method and results. This chapter is concluded with a discussion and implications.

Gender equality, driven by women's empowerment (Freeman & Svels, 2022), recognises the equal rights for fair treatment of all individuals, thus, benefiting society by allowing full and equal participation. The progress in achieving gender equality has been made (Bilan et al., 2020; Madsen & Scribner, 2017; Soare et al., 2022); however, there are still some existing aspects of gender disparities that require resolution.

The significant aspect of gender disparity is the gender wage gap (Amado et al., 2018; Coron, 2020; Lips, 2013), which refers to the consistently lower female earnings, compared to male, for performing similar tasks. What is the main concern about the gender wage gap is its persistence despite the fundamental principle in contemporary societies of providing equitable compensation for equivalent work (Amado et al., 2018). The digitalisation of an economy, as the main development trend, does not significantly reduce gender disparities in wages (Zhang, Tian, et al., 2023); in the ICT sector the gender wage gap is approximately 19% (Di Vaio et al., 2023).

Another aspect of gender inequalities is related to women's participation in certain sectors or occupations. The primary labour markets are characterised by their favourable earnings, working conditions, career prospect, etc., compared to secondary sectors, with women being under-represented in primary sectors and over-represented in secondary ones (Damelang & Ebensperger, 2020; Jamali et al., 2008; Mora & Muro, 2015; Symeonaki & Filopoulou, 2017). A significant facet of gender disparity is female under-representation in influential roles such as senior management (Lewellyn & Muller-Kahle, 2020; Madsen & Scribner, 2017; Soare et al., 2022), parliamentary members (Alexander et al., 2016) and entrepreneurship (Gaweł & Mińska-Struzik, 2023; Hägg et al., 2022; Ughetto et al., 2020).

Entrepreneurship is a gendered phenomenon from the quantitative perspective, with men significantly more likely to establish and run their businesses than women (Mustafa & Treanor, 2022) – in the EU countries women on average account for 30% of entrepreneurs (Dilli & Westerhuis, 2018; Ughetto et al., 2020) – and from the qualitative perspective, as female-led businesses are mostly smaller and less profitable than male-led (Reichborn-Kjennerud & Svare, 2014). The cultural perception of stereotypical entrepreneurs associates them with masculinity, primarily that of a white adult male (Jones et al., 2019; Williams & Patterson, 2019).

The uniqueness of the situation of women entrepreneurs with their specific limitations is recognised, along with the need to look for incentives to enter entrepreneurship as a possible manner of gaining empowerment and professional independence (Xu et al., 2023). A significant heterogeneity between men and women is observed regarding the influence of financial literacy and digital skills on the likelihood of becoming an entrepreneur (Oggero et al., 2020). Female entrepreneurs are less likely to attract external financial capital as compared to male entrepreneurs, with the penalising effect of stigma from previous business failures (Morazzoni & Sy, 2022; Pistilli et al., 2022).

Typically, the internal and external factors impacting individual decision to become an entrepreneurship are recognised (Dileo & García Pereiro, 2019; Saunoris & Sajny, 2017), as well as necessity-driven and opportunity-driven factors (Angulo-Guerrero et al., 2017; Nikolaev et al., 2018; Reissová et al., 2020). In the context of women, predictors of female entrepreneurship are recognised among gender-neutral factors similar to those of all entrepreneurs (Holmén et al., 2011), as well as among female-specific factors that reflect their unique circumstances (Dutta & Mallick, 2018; Gaweł & Mroczek-Dąbrowska, 2022; Pérez-Pérez & Avilés-Hernández, 2016). Digitalisation, including intelligent transformation, can be reflected in these aspects of being a potential predictor of female entrepreneurship.

Digitalisation is causing a profound transformation in the business environment (Almansour, 2023; Bansal et al., 2023; Chatterjee et al., 2022). Therefore, digital transformation, including AI technology, impacts business strategies and policies enabling firms to leverage technology for competitiveness (Zhai et al., 2023). The strategic changes brought about by the digitalisation resulted in the notion of digital entrepreneurship, referring to all entrepreneurial activities transferred to the digital sphere (Kraus et al., 2018; Zhai et al., 2023). From this perspective, digitalisation can be treated as a potential external predictor of female entrepreneurship impacting both the market functioning and internal processes of companies.

Coined by computer scientist John McCarthy in 1954, AI pertains to the capacity of an information system to exhibit actions, learning, comprehension and perception (Mariani et al., 2023). AI comprises a range of cognitive technologies, encompassing, for example machine learning or natural language processing (Kuziemski & Misuraca, 2020). AI helps to optimise resources, streamline production, facilitate organisational change, simplify decision-making in uncertainty and reshape business models (Malodia et al., 2023; Zhang, Jin, et al., 2023). Intelligent transformation, which encompasses the spread of AI technology from developers to enterprises, entails adopting AI to change business models towards processes of self-learning, self-optimisation, self-configuration and self-diagnosis (Wang et al., 2022). AI applications mitigate business risks, enhance capabilities and improve efficiency of companies (Drydakis, 2022).

Not only is AI believed to influence organisational resources but it can also extend its positive, long-term benefits to employee's well-being, work–life balance and engagement (Almansour, 2023). The development of AI services is gradually displacing human service providers, resulting in the fusion of products and services (Baek et al., 2023). There are also some wider, societal aspects of the diffusion of AI. For example, through automation, control over physical space, necessary resources and information, AI can be used to enforce citizen control (Kuziemski & Misuraca, 2020).

As AI influences various aspects of life, there is an urgent need to comprehend the novel socio-economic landscapes resulting from its widespread integration (Kuziemski & Misuraca, 2020). This chapter addresses this need and undertakes the investigation on how AI affects female entrepreneurship. The literature review presents contrasting viewpoints on the impact of digitalisation and AI on female entrepreneurship. While the current gender digital divide presents a potential risk of exacerbating gender disparities in entrepreneurship, there is also a perspective that gender digital dividend could hinder the rapid dismantling of gender-related obstacles in entrepreneurship, thereby narrowing the gap.

Digital skills and competences are needed to take advantage of new technologies (Hidalgo et al., 2020). From this perspective, the existence of the gender digital divide (Sánchez-Rivero et al., 2023; Vimalkumar et al., 2021) is a concerning aspect of digitalisation and its impact on female entrepreneurship. The concept of the digital divide, referring to the unequal access to and utilisation of ICT, has developed across three tiers: the first level involves ICT connectivity, the second level encompasses skills and competencies in the use of ICT, and the third level gauges the outcomes arising from the utilisation of the internet (Hidalgo et al., 2020). These three aspects of digital divide are also analysed in the context of gender gap (Picatoste et al., 2023).

Women are recognised as having limited access to digital technologies and infrastructure (Pawluczuk et al., 2021; Yeganehfar et al., 2018). In the context of formal skills and competences, the under-representation of females in the ICT workforce, among Science, Technology, Engineering and Mathematics (STEM) graduates, and at the university-level ICT education is widely observed (Lechman & Popowska, 2022; Patterson et al., 2021; Sattari & Sandefur, 2019; Yeganehfar et al., 2018). According to the data from Eurostat, the statistical office of the European Union, in the EU countries, women account for approximately 20% of employed ICT specialists and students enroled in tertiary education.¹ Additionally, the gender pay gap in the ICT sector is also recognised (Di Vaio et al., 2023). The under-representation of females persists despite the absence of gender-related disparities in STEM academic achievements when considering STEM subject scores, implying rather the stereotypical nature of gender bias and discrimination (Ho et al., 2020). The lower level of digital knowledge and interests, as well as male domination, are the examples of challenges for women operating in digital economic activities (Islam et al., 2023). This aspect of gender disparities in digital competencies can be treated as the individual level predictor of female entrepreneurship.

A common stereotype of males being more capable to STEM (Ho et al., 2020) might also lead to the fact that, compared to women, men more frequently undertake tasks that require ‘skills of the future’ related to STEM or ICT (Egana-delSol et al., 2022). In addition, ICT specialists are more prevalent in companies with male managers (Sánchez-Rivero et al., 2023). The ICT start-ups demonstrate a limited range of diversity in their workplace, predominantly being founded by young, single, well-educated and middle-class white males (Ferratti et al., 2021).

A consequence of gender stereotypes in the ICT sector is unequal gender participation in the development, implementation and use of ICT (Oleksy et al., 2012; Young et al., 2023), which results in bias in data, algorithms and applications of AI (Allen & Masters, 2020; Lauterbach, 2019; Liang et al., 2021). As men are the mostly likely to get position of AI developers, not women, algorithms of AI are developed being biases towards male perspective, not taking into account female perspective (Cachat-Rosset & Klarsfeld, 2023). By doing so, AI algorithms are also known to cause bias and disadvantageous at the user's side (Liang et al., 2021; Lütz, 2022). In consequences, women are less likely to meet all criteria developed by AI applications when applying for jobs (Kim & Heo, 2022; Kshetri, 2021).

The explanation for the gender digital divide lies in gender stereotypes and the process of gender socialisation, which discourages women from engaging with digital technologies (Sánchez-Rivero et al., 2023). In this context, the importance of training to develop women's digital skills is highlighted in the discussion on bridging the digital divide (Zhang, Tian, et al., 2023), as a factor helping to overcome social stereotypes. In fact, in female entrepreneurship in digital transformation two stereotypes overlap: both of entrepreneurship (Jones et al., 2019; Williams & Patterson, 2019) and the ICT sector (Ho et al., 2020) as stereotypically associated with men. As the impact of entrepreneurs' digital self-efficacy on company digitalisation has been demonstrated (Malodia et al., 2023), the gender digital divide can diminish the digital self-efficacy of female entrepreneurs, consequently impeding the pace of digitalisation within their businesses. Gender-related digital stereotypes can serve as external factors that influence societal norms and their impact on female entrepreneurship.

Access to funding for implementing digital technologies is another obstacle to achieving successful digital and intelligent transformation (Abaddi & AL-Shboul, 2023; Islam et al., 2023). Since female entrepreneurs face fewer opportunities to attract external funds (Pistilli et al., 2022), this barrier can play a significant role in sustaining gender gap in entrepreneurship with the advancement of AI technologies.

Given that the absence of digital competences is proven to be a significant barrier in the integration of AI into the business operations (Wang et al., 2022), the gender digital divide can be an important obstacle for women when establishing and running their own business both directly, by limiting the intelligent transformation of female-led ventures, and indirectly, by diminishing their capacity to engage in collaborative efforts within the intelligent environment. In this context, the gender digital divide can reduce the level of female entrepreneurship.

However, according to the theory of occupational choice, entrepreneurship is sometimes perceived as an alternative occupational choice to wage employment (Capelleras et al., 2019; Pardo & Ruiz-Tagle, 2017). As AI advances, concerns about the future of employment are also on the rise (AlQershi et al., 2023; Egana-delSol et al., 2022). AI algorithms used in human resource management are often biased when job applications are analysed in a way that discriminates against women (Kim & Heo, 2022; Kshetri, 2021). Women face a higher risk of automation due to workplace digitalisation compared to men (Egana-delSol et al., 2022). Being excluded or at the risk of exclusion from the wage employment due to the gender digital divide can serve as a necessity-driven motivation to establish one's own company. This may increase the level of female entrepreneurship, particularly in traditional sectors rather than digital-intensive ones.

There are also findings and arguments suggesting that the process of digitalisation can contribute to promoting greater gender equality and a more unbiased approach to conducting business, thereby overcoming traditional gender barriers and acting as a gender digital dividend (Di Vaio et al., 2023; Gaweł & Mińska-Struzik, 2023; Leong et al., 2022; McAdam et al., 2020; Ughetto et al., 2020). It has been shown that women benefit more from digital transformation compared to men through an increase in employment opportunities for women (Aly, 2022). Digitalisation supports women in their professional careers through its impact on achieving gender equal culture in the workplace (Chen et al., 2021). A new business culture marked by the democratisation of access to resources and the overcoming of traditional entry barriers may increase the number of female entrepreneurs on the internet (Ughetto et al., 2020). There is also evidence that the digital readiness of an economy positively affects women's involvement in entrepreneurship (Moeini Gharagozloo et al., 2023).

The digitalisation of their businesses should support women in fulfiling their business and family roles through more flexible working hours, in reaching the market more easily, and lowering the costs of operations (Alhajri & Aloud, 2023). For example, when considering conversational AI, which enables seamless and natural interactions between humans and computers (Mariani et al., 2023), it can facilitate gender-neutral conversations unlike interpersonal interactions that might be influenced by gendered norms and stereotypes. Digitalisation, including blockchain technology, can support women in financial markets by eliminating the need for physical currency and enhancing access to credit and markets (Di Vaio et al., 2023).

The gender digital dividend can serve as an opportunity-driven motivation for women to enter entrepreneurship. Digitalisation, including the diffusion of AI technologies, can create new opportunities for women to run business more independently and in a more gender-neutral manner, thus gaining a competitive advantage. This may increase the level of female entrepreneurship.

With all these overlapping and often contradictory arguments in mind, the research aims to investigate the problem whether the diffusion of AI is another obstacle for women entering entrepreneurship or rather the enabler. The cluster analysis is conducted at the macroeconomic level using data from the EU countries. The choice of the EU member states is motivated by their shared institutional background. Despite their national diversity, the EU countries share common values and norms, rooted in the EU institutional environment.

The following 27 countries, all members of the European Union, are included in the research: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain and Sweden. The data encompassing female entrepreneurship and the adoption of the AI within companies is utilised to identify the clusters of countries that share similarities within the cluster and distinctions from other clusters in these aspects. All data used in the research are secondary data, sourced in an open-accessed European Statistical Office EUROSTAT database for the year 2021. The choice of the year of study (2021) is motivated by the availability of data on the use of AI in companies. All data used in the research, together with their descriptive statistics, are explained in Table 1.

Female entrepreneurship is understood in this research as the share of women among self-employed. In 2021, the EU national average shows that women account for about 32% of entrepreneurs, oscillating between the minimum value of 23.5% and the maximum value of 40.8%. The diversity of EU countries in the spread of business use of AI is even greater. On average, 8.1% of companies in the European Union use at least one AI technology, ranging from 1.4% to 23.9% depending on a country; while on average 5.9% of them use AI for at least one business purpose, varying between 1.1% and 16.4% depending on a country.

Among the AI technologies, AI technology aimed at robotic process automation and AI technology for written language analysis are most often used by companies operating in the EU countries. AI technology dedicated to automating processes is used by an average of 3.2% of companies in the European Union (depending on the country, it ranges from 0.7% to 16.9%), while AI text mining technology is used by 3% of companies (with differences between countries from 0.4% to 10%). The least widespread AI technology is technology that enables the physical movement of machines, which is used by an average of 1% of companies in the European Union, ranging from 0.1% to 3.6% depending on the country.

Looking at the business purposes, AI technology focused on ICT security is the most widespread, as on average 2.3% of EU companies use this technology, with differences between countries from 0.5% to 8%. The second most explored area of AI use is marketing and sales – 2.2% of companies in the European Union use AI technologies for this purpose. AI technology dedicated to human resources management is the least used, as on average 0.8% of EU companies implement AI in this area of operation, ranging from 0% to 3% of companies depending on the country.

The diversity of EU countries, both in terms of women's entrepreneurship and the diffusion of AI technologies, justifies grouping them into clusters with similar characteristics within each cluster and different from other clusters. Through the use of cluster analysis, different patterns of female entrepreneurship and the spread of AI are to be uncovered. The k-means clustering method is used, with the number of clusters determined by Ward's minimum variance technique, assuming the significance level of p < 0.05.

As there are data on three aspects of the AI technology diffusion (the general aspect of AI diffusion, the usage of diverse AI technologies and the use of AI for different business purposes), the cluster analyses are conducted on these three aspects. The first cluster analysis is grounded in data concerning female entrepreneurship and the overall extent of AI adoption within companies. The second cluster analysis is founded on data related to female entrepreneurship and detailed measures regarding AI technologies. The third clustering is based on data on female entrepreneurship and AI application for various business purposes.

In the first cluster analysis, three variables are taken to identify the clusters for the EU countries: female entrepreneurship (FEE), the share of companies using at least one AI technology (AI one tech) and the share of companies using AI technology for at least one purpose (AI one purpose). Based on these three variables, while using the k-means clustering method (Table 2), three clusters of EU countries are identified.

Looking at the descriptive statistics of these clusters (Table 3), cluster# 1 is characterised by the highest level of female entrepreneurship (on average in countries belonging to cluster# 1, women account for 36.4% of entrepreneurs), while cluster# 2 by the lowest (28.2% of entrepreneurs are women). The level of AI diffusion is relatively low, significantly lower than in cluster# 3, with, on average, 6.7% of companies using at least one AI technology and 4.4% using at least one business purpose in cluster# 1, and with 5.7% of companies using at least one AI technology and 4.2% using at least one business purpose in cluster# 2. The last identified cluster# 3 is characterised by a moderate level of female entrepreneurship compared to other two clusters (women account for 33.5% of entrepreneurs) and the highest level of the AI diffusion, as, on average, 16.4% of companies in the countries belonging to cluster# 3 utilise at least one AI technology and 13% of them explore AI technology for at least one business purpose (Table 4).

In the second approach to cluster research, i.e. identifying the clusters, data on female entrepreneurship (FEE) and on the diffusion of various AI technologies among enterprises operating in the EU countries is used. Specifically, the following are considered: AI technologies for the analysis of written language (AI text mining), for converting spoken language into machine-readable format (AI speech recognition), for generating written or spoken language (AI natural language generation), for identifying objects or persons based on images (AI images recognition), for machine learning for data analysis (AI machine learning for analysis), for automating different workflows or assisting in decision-making (AI robotic process automation) and for enabling the physical movement of machines (AI physical movement). While using the k-means clustering method, the variable presenting the diffusion of AI technology identifying objects or persons based on images (AI images recognition) proved to be statistically significant in grouping countries. Thus, it was withdrawn and the k-means clustering method was repeated with one variable less. This allowed us to identify three clusters of EU countries (Table 4).

The descriptive statistics on FEE and on the diffusion of different AI technologies among enterprises operating in the EU countries in the identified clusters are presented in Table 5. Cluster# 1 is characterised by the highest representation of women among entrepreneurs (on average 36.6%), and relatively low levels of all AI technologies' diffusion, and the lowest level of diffusion of AI technologies for generating written or spoken language. Cluster# 2 is the one with the lowest participation of females in entrepreneurship (on average, women account for 28.3% of entrepreneurs in the countries belonging to this cluster), and with the lowest diffusion of almost all AI technologies (0.6%–2.1% of companies using various AI technologies). In countries belonging to cluster# 3, the level of spread of all AI technologies is the highest (from 2.9% to 12.8% of companies in countries belonging to this cluster use AI technologies, depending on the type of technology). At the same time, the level of female entrepreneurship in the countries in this cluster is moderate; on average, 31.9% of entrepreneurs are women.

In the last clustering, FEE is considered together with the diffusion of various purposes of the business use of AI technologies', including the use for marketing or sales, production processes, business administration processes, management, logistics, ICT security and human resource management. The introduction of k-means cluster method allows to identify three clusters of EU countries (Table 6).

The descriptive statistics of the clusters (Table 7) show the diversity of the clusters. In the countries belonging to cluster# 1, the level of female entrepreneurship is the highest, as, on average, women account for 36.4% of entrepreneurs. However, the level of use of AI technologies for different business purposes is relatively low here, ranging from 0.6% to 1.7% of companies exploring AI technologies, depending on the business purpose. Cluster# 2 is also characterised by a low level of prevalence of the use of AI technologies for business purposes, as between 0.5% and 2.2% of companies use them, depending on the purpose. In this cluster, the level of female entrepreneurship is the lowest, as women represent 28.3% of entrepreneurs. In cluster# 3, the spread of AI technology use for business purposes is the highest, with 2.2%–5.4% of companies using AI technology depending on the purpose. This cluster is also characterised by moderate levels of female entrepreneurship, accounting for 34.6% of entrepreneurs.

The cluster analysis shows a relationship between the female entrepreneurship level and the diffusion of AI technology in business, with an inverted U-shaped pattern. Three clusters of the EU countries are identified in all three research attitudes. Regardless of the attitude towards the measures of AI diffusion, the similar patterns are discovered. A relatively low level of diffusion of AI technology occurs in two clusters of countries, where one of these clusters is characterised by a relatively low level of female entrepreneurship, while the other cluster has a relatively high level of female entrepreneurship. The highest level of the utilisation of AI technology is observed in the third cluster of countries, exhibiting a moderate level of female entrepreneurship.

The discovery of an inverted U-shaped relationship between female entrepreneurship and the diffusion of AI technology in business sheds new light on the previously conflicting results. Both concerns rooted in the gender digital divide (Lechman & Popowska, 2022; Patterson et al., 2021; Pawluczuk et al., 2021; Sánchez-Rivero et al., 2023; Sattari & Sandefur, 2019; Vimalkumar et al., 2021; Yeganehfar et al., 2018), regarding the possible deepening of the gender gap in entrepreneurship within the context of digital and intelligent transformation, and the belief in the gender equalising effect of digitalisation (Aly, 2022; Chen et al., 2021; Di Vaio et al., 2023; Leong et al., 2022; McAdam et al., 2020; Ughetto et al., 2020), might be correct, as it appears that this relationship is multidimensional and dependent on the level of AI diffusion.

There is a need for diverse policies and practices due to the variety of patterns in female entrepreneurship and the spread of AI technology in EU countries. The integration of the female perspective into the diffusion of AI is needed to explore the AI technology for the sake of a more sustainable society. Risk factors that could potentially exclude women from entrepreneurship due to AI technology exist and must be pre-emptively addressed as AI continues to advance. Nonetheless, there are instances of countries successfully managing both the implementation of AI and the attraction of women to entrepreneurship.

The results of the research contribute to a better understanding of intelligent transformation and its societal impacts. The inverted U-shaped pattern of the interrelationship between the spread of AI technologies in business and female entrepreneurship sheds light on the multidimensional and country-specific nature of the diffusion of AI, and, most likely, of digitalisation as well. Given that the progress of intelligent transformation can lead to both the deepening and equalising of the gender gap in entrepreneurship, there is a need for social awareness and efforts to ensure that intelligent transformation benefits all, fostering more equal and just societies. The results of clustering indicate the need of country-specific actions supporting different social groups in the digital transformation, as recommended by Egana-delSol et al. (2022).

The results also contribute to the theory of entrepreneurship, particularly in explaining the gender gap in entrepreneurship. Women entrepreneurs are embedded within the social and business context, influenced by external factors in the entrepreneurial environment. The importance of the diffusion of AI technology as an external factor resulting in women's engagement in entrepreneurship is indicated. The results of cluster analyses show the diverse responses of women entrepreneurs to the diffusion of AI, advocating for a more contextual and country-specific understanding of the impact of external factors on women's decisions to become entrepreneurs.

To fully benefit from intelligent transformation, the digital divide should be reduced. In line with the recommendation for specific policy action to reduce the digital divide (Hidalgo et al., 2020), the results imply the need for training and education programmes dedicated to female entrepreneurs and to all entrepreneurs with a lower level of digital competences. Training programmes should focus on empowering under-represented groups with limited digital exposure, such as female entrepreneurs, by enhancing their digital skills and competence in utilising AI technologies for their businesses. These programmes should aim not only to raise the level of digital knowledge and skills but also to address and overcome gender digital stereotypes.

Other ways to reduce the gender digital divide and overcome gender-related obstacles in intelligent transformation can involve providing consulting and implementation services tailored to the utilisation of AI technology for female entrepreneurs and all entrepreneurs with lower levels of digital competence. Promoting knowledge and awareness of the possibilities of using AI in business, especially for female entrepreneurs and those with lower levels of digital competence, can also be supportive. An inclusive, easy-to-implement and user-friendly AI application can help reduce competency barriers in the adoption of AI as well.