Mind the AI Gap: Asymmetrical Age Differences in Entrepreneurs’ Perceptions of Artificial Intelligence

Abstract

As artificial intelligence (AI) becomes embedded in entrepreneurial practice, an unresolved question is whether age shapes founders’ perceptions of its opportunities and risks. Drawing on diffusion-of-innovations and technology adoption theories, this study examines whether age cohorts differ in their perceived benefits of AI, perceived risks, and short-term expectations regarding AI’s business impact. Using data from the 2024 Global Entrepreneurship Monitor (GEM) survey for Slovenia, we analyze ordinal indicators across all three domains. Bivariate comparisons using Mann–Whitney U tests with effect sizes are complemented by multivariate ordinal logistic regression models controlling for sector, education, and gender. The analysis reveals an asymmetrical age gap in AI perceptions. Younger entrepreneurs report higher perceived benefits and more positive impact expectations, while AI-related risk perceptions do not vary by age. Multivariate analyses show that age effects on perceived benefits are context-dependent, whereas age remains a robust predictor of future-oriented impact expectations. The study offers a theoretically grounded and methodologically transparent analysis integrating technology adoption frameworks with entrepreneurial psychology. Practically, it underscores the need for differentiated AI-readiness initiatives that address age-related differences in strategic orientation and preparedness. Future research could further explore the roles of capabilities, industry context, and entrepreneurial experience.

1. Introduction

Artificial intelligence (AI) is increasingly integrated into entrepreneurial processes worldwide, including in smaller and emerging economies such as Slovenia, shaping how new ventures discover, design, and deliver value. For early-stage entrepreneurs, AI enhances opportunity recognition and market sensing, supports automation and optimization across business functions, and enables data-driven decision-making that shortens experimentation cycles and reduces resource requirements (Chalmers et al., 2021; Giuggioli & Pellegrini, 2023). Recent research positions AI as a general-purpose technology capable of transforming entrepreneurial ecosystems through new capabilities, competitive advantages, and business models (Iansiti & Lakhani, 2020). However, adoption outcomes remain uneven, depending on complementary digital skills, data readiness, and organizational capabilities (Faruque et al., 2024; Fossen et al., 2024; Truong et al., 2023).

While AI diffusion has accelerated globally, younger entrepreneurs are often found to display greater openness and willingness to experiment with emerging technologies than older founders (OECD/EC, 2020). Prior research suggests that younger entrepreneurs tend to be more digitally fluent, adaptive, and innovation-oriented, traits that support opportunity recognition and early experimentation with new technologies (Calvo-Porral & Pesqueira-Sanchez, 2020; Choudhary et al., 2024; Lewis & Massey, 2003). However, access barriers, such as limited resources, expertise, and strategic guidance, can restrict effective implementation across age groups, resulting in substantial heterogeneity in realized performance outcomes (de Klerk et al., 2025; Faruque et al., 2024).

Theoretical perspectives offer mixed expectations regarding how age shapes technology-related perceptions. Diffusion-of-innovations theory (Rogers, 2003) and technology adoption frameworks (Davis, 1989; Venkatesh et al., 2003, 2012) indicate that differences in digital fluency, switching costs, and facilitating conditions may affect perceived usefulness and ease of use, potentially leading younger founders to evaluate the benefits and future impact of AI more positively. In contrast, research on risk perception highlights shared evaluative baselines among populations exposed to similar informational and regulatory environments, which may limit age-based divergence in perceived risks (Slovic, 2016). Life-course theories add that evolving time horizons and experience may moderate investment expectations without fundamentally altering underlying concerns (Carstensen et al., 1999).

Despite growing interest in AI and entrepreneurship, empirical research rarely distinguishes between opportunity-oriented evaluations (such as perceived benefits and opportunity potential) and risk-related perceptions (such as bias, reliability, and cost) within the same comparative framework. It also does not systematically assess how these distinct evaluative dimensions differ between younger and older entrepreneurs. This omission limits understanding of whether observed age-related differences reflect variation in opportunity-oriented outlooks, differential exposure to AI-relevant contexts, or genuine differences in risk perception.

Building on these theoretical foundations, this study examines whether entrepreneurs’ age cohorts differ in their (i) perceived benefits of AI, (ii) perceived risks, and (iii) short-term expectations regarding AI’s business impact. Using data from the 2024 Global Entrepreneurship Monitor (GEM) Slovenia survey, the article contributes by (1) providing a cohort-based analysis that empirically distinguishes between opportunity-oriented evaluations (as reflected in perceived benefits of AI) and concern (risk perception) as independent evaluative dimensions, (2) clarifying when and how age is relevant in shaping technology-related perceptions, and (3) deriving practical implications for capability-building initiatives and AI readiness programs targeting entrepreneurs at different life stages. While the empirical focus is on Slovenia, the study’s theoretical contribution is relevant to broader entrepreneurial contexts, offering insight into how age-related cognitive, motivational, and contextual factors shape perceptions of AI in comparable innovation ecosystems.

In this study, the term optimism is used exclusively in a theoretical sense, referring to a general opportunity-oriented evaluative disposition discussed in the literature on technology adoption and entrepreneurial psychology. Empirically, the analysis relies on indicators of perceived benefits of AI, which capture entrepreneurs’ assessments of AI’s positive effects on business functions. To maintain conceptual clarity, the term perceived benefits of AI is used consistently when referring to the measured variables. Despite the growing integration of AI into entrepreneurial practice, existing research provides limited insight into whether and how founders from different age cohorts interpret the opportunities and risks associated with AI. While many studies examine technology adoption or entrepreneurial attitudes more broadly, they rarely differentiate between opportunity-oriented evaluations, such as perceived benefits and opportunity potential, and concern-related perceptions, including risks, ethical issues, or implementation challenges, within a generational comparative framework. This gap is important because age is closely related to differences in digital fluency, motivational orientation, and perceived time horizons, all of which influence how entrepreneurs evaluate emerging technologies. A clearer understanding of these generational differences is therefore valuable for theory, as it helps identify whether variations in AI perceptions stem from cognitive, motivational, or life-course mechanisms, or from differential exposure to AI-relevant context. It is also relevant for practice, as it enables policymakers and support institutions to design AI-readiness initiatives that address the specific needs of both younger and older entrepreneurs. Addressing this gap, the present study provides an empirical comparison of age-based differences in perceived benefits of AI, perceived risks, and short-term expectations regarding AI’s business impact. Importantly, this study does not assume a uniform generational gap across all dimensions of AI perception. Instead, it explicitly examines the possibility of an asymmetrical gap, in which age-related differences are more likely to emerge in opportunity-oriented evaluations of AI, while perceptions of AI-related risks remain broadly shared.

Section 2 reviews the relevant literature on AI in the entrepreneurial process and theoretical perspectives on technology perception and age. Section 3 outlines the research design, data, and both bivariate and multivariate analytical procedures. Section 4 presents the results of the comparative analysis between younger and older entrepreneurs, while Section 5 discusses the theoretical and practical implications of the findings. Section 6 concludes with a summary of key insights, limitations, and directions for future research.

2. Literature Review and Theoretical Background

2.1. AI in the Entrepreneurial Process: Capabilities and Effects

Digital transformation reconfigures the architecture of value creation in entrepreneurship. Beyond streamlining operations, it entails the strategic redesign of business models and organizing logics (Verhoef et al., 2021; Vial, 2021). AI extends this trajectory. By enabling data-intensive sensing, prediction, and automation across the venture lifecycle, AI operates as a general-purpose capability that alters how new ventures discover opportunities, experiment with solutions, and deliver value at scale (Nambisan et al., 2019).

Within opportunity recognition, AI enhances market sensing and ideation by extracting patterns from heterogeneous data (search trends, customer reviews, social media, and user telemetry), supporting faster hypothesis generation, early validation, and continuous learning loops (Chalmers et al., 2021). These capabilities align with broader evidence that AI acts as a catalyst for dynamic business model innovation and value creation processes (Jorzik et al., 2024). During venture creation and product development, learning systems and generative models accelerate experimentation and prototyping, improve quality assurance, and compress build–measure–learn cycles central to agile development (Iansiti & Lakhani, 2020; McKinsey & Company, 2023). In commercialization and operations, AI supports personalization in marketing and sales, forecasting in supply chains, financial decision-making, and service automation—enhancing workflow efficiency and cost optimization (Giuggioli & Pellegrini, 2023).

Reported outcomes of digitalization include enhanced entrepreneurial activity, improved productivity, and stronger competitiveness, reflecting how digital integration and technological transformation foster innovation and performance at both firm and ecosystem levels (Dabbous et al., 2023; del Olmo-García et al., 2023; OECD, 2024). Impact, however, is conditional rather than automatic. Prior work consistently highlights the role of complementary assets—founders’ digital skills, data readiness (availability, quality, governance), and change-management practices, in translating adoption into realized performance (OECD/EC, 2020; Verhoef et al., 2021; Vial, 2021).

A salient feature of the current diffusion is the democratization of access: cloud-based services, open-source frameworks, and low- or no-code interfaces have significantly lowered fixed costs and extended advanced AI capabilities to startups and SMEs (McKinsey & Company, 2023; OECD/BCG/INSEAD, 2025). This ‘leveling of the technological playing field’ expands experimentation under resource constraints, yet persistent disparities in digital literacy, data quality, and organizational adaptation continue to separate pilots from scalable integration (Kahveci, 2025; Silva et al., 2022).

Two implications follow for our research design. First, AI capabilities extend across multiple entrepreneurial functions—spanning opportunity recognition, product development, commercialization, and operations—indicating that positive opportunity-oriented evaluations of AI are inherently multidimensional rather than unidimensional attitudes (Jorzik et al., 2024; Nambisan et al., 2019). Second, perceptions of AI-related risks (such as bias, reliability, privacy and data security, or workforce displacement) are increasingly shaped and standardized through public discourse, regulatory initiatives, and professional ethics frameworks. These shared narratives and institutional norms help homogenize how potential downsides are appraised, even as opportunity-oriented evaluations of AI continue to vary partly across cohorts differing in digital fluency and experience (Burrell, 2016; Wirtz et al., 2019). These insights motivate our cohort-based comparison of perceived benefits, perceived risks, and expected business impact.

2.2. Adoption Frameworks: From Technology Diffusion to Entrepreneurial Decision-Making

Understanding why entrepreneurs differ in their perceptions of AI’s benefits and risks requires grounding in established theories of technology adoption. These frameworks share a common premise: adoption is determined not only by technological potential, but also by how individuals and organizations perceive, evaluate, and contextualize emerging innovations.

The Diffusion of Innovations (DOI) theory (Rogers, 2003) conceptualizes adoption as a social process through which new ideas and technologies spread over time. Adoption decisions depend on five perceived attributes of an innovation: relative advantage, compatibility, complexity, trialability, and observability, which together shape when and how individuals adopt. For entrepreneurs, these perceptions are influenced by prior experience, learning orientation, and openness to change. Younger founders, who matured in digitally pervasive environments, are expected to perceive AI as a natural extension of existing infrastructures, emphasizing opportunity and experimentation. In contrast, older founders may place greater emphasis on reliability, controllability, and proven outcomes, reflecting broader generational patterns observed in technology adaptation and readiness (Dwivedi et al., 2020; Parasuraman & Colby, 2015).

At the individual, cognitive level, the Technology Acceptance Model (TAM) (Davis, 1989) explains behavioral intention through two core beliefs: perceived usefulness and perceived ease of use. Its unified successors, UTAUT and UTAUT2 (Venkatesh et al., 2003, 2012), refine these mechanisms by incorporating social influence, facilitating conditions, and hedonic motivation. In entrepreneurial settings, these constructs translate into judgments about how AI enhances opportunity recognition, decision quality, and operational efficiency. Differences in digital self-efficacy and technology readiness, often associated with age, shape evaluations of AI’s usefulness and expected effort and thus influence entrepreneurs’ perceptions of AI adoption (Parasuraman & Colby, 2015).

The Technology Readiness Index 2.0 (TRI) further distinguishes between positive dispositions (optimism and innovativeness) and negative ones (discomfort and insecurity) that jointly influence adoption behavior. In entrepreneurship, TRI-based optimism aligns with a proactive view of AI as a creative and empowering enabler that enhances innovation and problem-solving. Conversely, discomfort reflects anxiety over automation, potential loss of human autonomy, and increasing opacity in algorithmic decision-making (Haenlein & Kaplan, 2019; Parasuraman & Colby, 2015). These contrasting dispositions capture the dual nature of technological readiness. While optimism in the TRI sense reflects a general positive disposition toward technology and is theoretically associated with viewing AI as a strategic opportunity for experimentation and competitive advantage, discomfort captures concerns related to ethics, control, and dependency. Such ambivalence is particularly salient in entrepreneurial contexts where technology both expands and destabilizes traditional boundaries of human agency and organizational control.

Beyond individual cognition, adoption is embedded in broader organizational and institutional contexts. The Technology–Organization–Environment (TOE) framework (Tornatzky et al., 1990) posits that technological readiness, organizational capability, and external environmental pressures jointly shape an organization’s propensity to adopt and implement innovations. Age and experience moderate these forces. Older entrepreneurs may rely on established routines and social capital to navigate uncertainty, while younger entrepreneurs may compensate through agility, experimentation, and a learning orientation. As AI shifts from optional to strategic infrastructure, voluntariness of use declines, and adoption becomes increasingly influenced by normative, competitive, and regulatory expectations (Haenlein & Kaplan, 2019).

While these adoption models illuminate the cognitive and structural determinants of technology uptake, understanding entrepreneurs’ reactions to AI also requires attention to the affective and motivational factors underlying risk and opportunity perception.

2.3. Risk Perception and Entrepreneurial Psychology

Entrepreneurship inherently involves operating under uncertainty, where decisions must balance opportunity recognition against risk management. Entrepreneurs’ perceptions and evaluations of emerging technologies such as AI depend not only on their cognitive understanding of usefulness but also on their affective and motivational orientations. Research in entrepreneurial psychology consistently shows that risk perception is subjective, shaped by experience, emotion, and heuristics rather than by objective probabilities (Keh et al., 2002; Simon et al., 2000).

Psychological dispositions such as trait-level optimism and risk perception play central roles in this process. Entrepreneurial optimism, a generalized expectation of favorable outcomes, serves as a psychological resource that sustains motivation and creativity in uncertain contexts. Hmieleski and Baron (2009) show that such optimism fosters persistence and adaptive behavior, helping entrepreneurs remain proactive despite setbacks. Extending this view, Cassar (2010) finds that optimistic expectations, though sometimes overly positive, encourage individuals to pursue new ventures and commit to exploiting opportunities even under risk. However, trait-level optimism is not evenly distributed across individuals. Drawing on socioemotional selectivity theory, Carstensen et al. (1999) argue that younger individuals, perceiving a longer future time horizon, tend to be more open to experience and tolerant of risk, while older adults prioritize emotional stability and realistic goals. Similarly, Levesque and Minniti (2006) find that as entrepreneurs age, shorter investment horizons and accumulated experience lead to more measured optimism and cautious opportunity evaluation. According to Socioemotional Selectivity Theory (Carstensen et al., 1999), as individuals age and perceive their future time horizons as shorter, motivational priorities shift from exploration and growth to emotional meaning and stability.

The broader literature on risk perception emphasizes that individuals evaluate risk through two intertwined processes: analytic and affective (Slovic, 2016; Weber et al., 2002). Analytic evaluation is deliberate and data-driven, relying on probabilistic reasoning and factual assessment of potential outcomes. In contrast, affective evaluation operates intuitively. People respond to perceived threats and opportunities through emotional cues shaped by past experiences, social narratives, and cultural meanings, a mechanism often described as “risk-as-feeling.” Because such affective responses are socially transmitted and reinforced through media, policy discourse, and collective experience, they tend to converge across populations over time (Slovic, 2016). This dynamic helps explain why perceptions of AI-related risks, such as bias, privacy concerns, or job displacement, may show less variation across age groups than expectations regarding AI’s potential benefits and near-term business impact.

Entrepreneurship research further emphasizes the importance of regulatory focus for understanding entrepreneurial behavior. Promotion-oriented entrepreneurs seek advancement, experimentation, and innovation, while prevention-oriented entrepreneurs prioritize security, control, and the avoidance of failure (Brockner et al., 2004; Higgins, 1997; Hmieleski & Baron, 2009). Age influences this motivational orientation. Younger founders tend to be more exploration- and growth-oriented, reflecting a promotion focus, while older entrepreneurs often value stability and control, consistent with a prevention focus (Kautonen et al., 2015; Lockwood et al., 2005). These motivational orientations also shape how entrepreneurs interpret emerging technologies. A promotion focus aligns with viewing AI as an enabler of efficiency and innovation (Huang & Rust, 2018), while a prevention focus heightens attention to ethical, security, and dependency risks (Wirtz et al., 2019).

Collectively, this perspective places age-related differences in AI perception within broader psychological mechanisms of motivation and risk appraisal. Motivational orientations influence how entrepreneurs balance perceived benefits and risks, shaping their readiness to integrate AI into business practice. Understanding these dynamics helps explain generational variation in technology adoption and informs support programs tailored to different motivational profiles (Hmieleski & Baron, 2009; Huang & Rust, 2018).

Taken together, the reviewed frameworks highlight the multifaceted nature of age effects on entrepreneurs’ perceptions of artificial intelligence. Diffusion-of-innovation and adoption models explain cognitive evaluations of AI’s usefulness and ease of use, predicting higher benefit perceptions (i.e., opportunity-oriented evaluations of AI), which are theoretically related to, but empirically distinct from, psychological optimism discussed in the entrepreneurship literature. Psychological and motivational theories add an affective layer, indicating that optimism in a psychological sense and prevention motives diverge across life stages: younger entrepreneurs adopt a promotion-focused orientation, while older entrepreneurs emphasize stability and control. Organizational perspectives such as the TOE framework provide the contextual dimension, showing that experience and routines shape how these perceptions translate into adoption readiness.

Rather than proposing a causal or moderating model, this study adopts a comparative perspective grounded in technology adoption and entrepreneurial psychology theories to examine age-related differences in entrepreneurs’ perceptions of AI.

Specifically, it addresses the following research questions:

• RQ1: Do younger and older entrepreneurs differ in their perceptions of the benefits of AI?
• RQ2: Do younger and older entrepreneurs differ in their perceptions of the risks associated with AI?
• RQ3: Do younger and older entrepreneurs express different expectations regarding the future impact of AI on their businesses?

To anchor the research questions in relevant theoretical mechanisms, it is important to recognize that younger entrepreneurs are often characterized by higher levels of digital fluency, longer perceived time horizons, and a promotion-focused motivational orientation. These characteristics are commonly associated with stronger expectations regarding emerging technologies. Older entrepreneurs, by contrast, may draw more heavily on accumulated experience and shorter perceived time horizons, which may shape more cautious or deliberative evaluations of technological change. Taken together, these mechanisms suggest that age-related differences are more likely to manifest in perceptions of AI’s benefits and expected business impact than in perceptions of AI-related risks, thereby providing a theoretical rationale for the research questions guiding this study. By addressing these research questions, the study aims to clarify whether observed generational differences arise from variation in opportunity-oriented evaluations of AI, perceptions of risk, or expectations regarding technological payoff. In doing so, it links cognitive and motivational theories of technology adoption with empirical patterns in entrepreneurship and contributes to a more nuanced understanding of how entrepreneurs of different ages perceive AI.

3. Research Design and Methodology

3.1. Data and Variables

The empirical analysis is based on data from the Global Entrepreneurship Monitor (GEM) Slovenia 2024 database. GEM provides harmonized, internationally comparable data on entrepreneurial attitudes, activities, and ecosystem conditions through the Adult Population Survey (APS).

The Slovenian 2024 dataset includes approximately 2000 adult respondents, representing the national population aged 18–64 years. For the purpose of this study, we focus on early-stage entrepreneurs, defined according to GEM methodology as individuals actively engaged in founding a business or managing one less than 42 months old. This entrepreneurial subsample represents roughly 6–8% of the total survey population.

The subsample reflects respondents at the formative stages of entrepreneurial activity, where openness to innovation and the adoption of new technologies are particularly relevant. The approach follows standard GEM sampling and weighting procedures, ensuring the reliability and comparability of results across countries and over time. After applying TEA criteria, the subsample consisted of N = 136 respondents. Although the resulting subsample is relatively small, it is appropriate for the exploratory and comparative analyses conducted in this study, which rely on ordinal indicators and nonparametric and ordinal regression methods.

3.2. Sample Characteristics and Measures

Table 1. Demographic data.

Characteristic	Category	n	%
Gender (N = 136)	Male	94	69.1
	Female	42	30.9
Age (N = 135)	Mean (SD)	40.24 (12.85)	—
	Range	18–74	—
Educational Attainment (N = 136)	None	4	2.9
	Some secondary	6	4.4
	Secondary degree	49	36.0
	Post-secondary	62	45.6
	Graduate level (Master’s/PhD)	15	11.0
Business Sector (N = 136)	Extractive	4	2.9
	Transforming	30	22.1
	Business services	60	44.1
	Consumer-oriented	42	30.9
Firm Size (Employees) (Valid N = 64)	No employees	46	71.9
	1–5 employees	16	25.0
	6–19 employees	2	3.1

summarizes the demographic and contextual characteristics of early-stage entrepreneurs. The mean age was 40.24 years (SD = 12.85; range: 18–74). Male entrepreneurs represented 69.1% of the sample. Educational attainment was generally moderate to high, with 45.6% reporting post-secondary education and 11.0% graduate-level qualifications. The majority operated in business services (44.1%) or consumer-oriented activities (30.9%). Firm size was typically very small: among valid responses, 71.9% had no employees and 25% employed 1–5 workers, reflecting the early-stage nature of TEA ventures.

Within this subsample, respondents were divided into two age-based cohorts, following established conventions in entrepreneurship and life-course research:

• Younger entrepreneurs (18–34 years), often associated in prior research with higher digital fluency, greater flexibility, and a stronger tendency toward opportunity-oriented evaluations of new technologies;
• Older entrepreneurs (35+ years), representing more experienced founders with accumulated business knowledge and established decision patterns (Levesque & Minniti, 2006).

The final sample included 36.3% younger entrepreneurs and 63.7% older entrepreneurs. One case was excluded due to missing age information. This distribution reflects the typical age structure of the early-stage entrepreneurial population in Slovenia, where mid-career and later-career founders constitute the majority of new business starters. The analysis focuses on three perception domains of artificial intelligence (AI), covering a total of 11 indicators, all measured on three-point Likert-type scales (1 = disagree, 2 = neither agree nor disagree, 3 = agree).

(1)

Perceived benefits of AI

Five items assessed perceived positive impacts of AI: customer personalization, productivity and operational efficiency, innovation in products and services, risk management and compliance, and business growth and revenue increase.

(2)

Perceived risks of AI

Five items reflected respondents’ concerns related to AI integration: data security and privacy, implementation costs and complexity, employee resistance, ethical dilemmas, and customer distrust.

(3)

Expected business impact of AI

Expected business impact was measured using a single ordinal item. While suitable for capturing short-term strategic expectations, this measurement limits reliability and does not permit the estimation of a latent construct; accordingly, the results should be interpreted as indicative rather than definitive.

Although the statistical tests were conducted at the item level, internal consistency checks indicated acceptable reliability for both benefit-related items and risk-related items (Cronbach’s α > 0.70).

3.3. Descriptive Statistics of Main Variables and Analytical Approach

Table 2. Descriptive statistics for AI-related perception indicators.

Indicator	N	Min	Max	Mean	SD
Perceived Benefits
Customer personalization	135	1	3	2.03	0.83
Productivity and efficiency	135	1	3	2.24	0.86
Product and service innovation	135	1	3	2.06	0.84
Risk management and compliance	134	1	3	1.90	0.79
Business growth potential	135	1	3	2.07	0.81
Perceived Risks
Data security and privacy	133	1	3	2.13	0.87
Implementation cost and complexity	131	1	3	1.73	0.73
Employee resistance	132	1	3	1.63	0.72
Ethical concerns	128	1	3	1.97	0.82
Customer distrust	133	1	3	1.85	0.80
Expected Business Impact (3 years from now)	136	1	3	1.92	0.79

Note: Due to item-level missing responses, valid Ns vary slightly across indicators (N = 128–136).

reports the descriptive statistics for all AI-related indicators. Perceived benefits showed generally neutral-to-positive mean values (1.90–2.24), whereas perceived risks were moderate (1.63–2.13), with data security and privacy representing the most salient concern. Expected near-term business impact was moderately positive (M = 1.92). Together, these descriptive patterns indicate balanced but moderately positive evaluations of AI.

Given the ordinal nature of the data and deviations from normal distribution, nonparametric methods were employed. The Mann–Whitney U test was used to examine statistically significant differences between the two age cohorts. This test compares ranked distributions and median values, offering a robust approach for moderately sized and unbalanced samples. The Mann–Whitney U test compares ranked distributions and median values, offering a robust approach for moderately sized and unbalanced samples. It is appropriate for independent samples when the assumptions of normality and homoscedasticity cannot be met, providing a robust inference framework for ordinal and non-normally distributed variables (Field, 2013). Prior to testing, data distributions and response patterns were inspected to ensure internal consistency and absence of systematic missingness.

To complement significance testing, effect sizes (r) were calculated using the formula:

r = |Z|/√N

(1)

allowing interpretation of the magnitude of group differences according to Cohen’s (1988) guidelines:

• r ≈ 0.10 → small effect;
• r ≈ 0.30 → medium effect;
• r ≥ 0.50 → large effect.

To assess whether observed age-related differences persist after accounting for potential confounding factors, additional multivariate analyses were conducted using ordinal logistic regression models. These models included age cohort as the focal predictor and controlled for gender, educational attainment, and sector affiliation. Ordinal logistic regression was selected due to the ordinal structure of the dependent variables and its suitability for moderately sized samples.

All analyses were conducted using IBM SPSS Statistics 29.0, with a significance threshold of p < 0.05.

The chosen research design operationalizes the conceptual framework by empirically examining whether younger and older entrepreneurs differ in (i) their perceptions of AI’s benefits, (ii) their perceptions of AI-related risks, and (iii) their expectations regarding AI’s short-term business impact. This framing reflects the study’s approach of treating opportunity-oriented evaluations of AI and risk-oriented evaluations as distinct empirical dimensions, while conceptual discussions of technological optimism are grounded in the relevant theoretical literature rather than in the measurement model. While the GEM dataset ensures methodological rigor and cross-national comparability, several limitations should be acknowledged. The analysis focuses on a specific subgroup of entrepreneurs within Slovenia, which may constrain generalizability. Moreover, the cross-sectional nature of the data captures perceptions at one point in time, potentially overlooking rapid changes in technological attitudes. Although the study combines bivariate nonparametric comparisons with multivariate ordinal regression models to account for key confounding factors such as gender, education, and sector affiliation, the relatively small sample size limits the inclusion of additional controls and the estimation of more complex interaction effects. Nevertheless, the standardized GEM methodology provides a robust empirical foundation for exploring generational differences in AI-related perceptions among entrepreneurs.

4. Results

The comparative analysis examined whether younger (18–34 years) and older (35+ years) entrepreneurs differ in their perceptions of artificial intelligence (AI) across three domains: perceived benefits, perceived risks, and expected business impact.

Table 3. Differences in perceptions of AI between younger and older entrepreneurs.

Domain	Indicator	Mann–Whitney U	p-Value	Effect Size (r)	Direction of Difference
Perceived Benefits	Customer personalization	1479.0	0.002	0.27	Higher (younger)
	Productivity and efficiency	1446.5	0.001	0.29	Higher (younger)
	Product and service innovation	1481.0	0.002	0.27	Higher (younger)
	Risk management and compliance	1659.0	0.031	0.19	Higher (younger)
	Business growth potential	1470.0	0.002	0.27	Higher (younger)
Perceived Risks	Data security and privacy	1866.5	0.297	0.09	n.s.
	Implementation costs and complexity	1979.5	0.879	0.01	n.s.
	Employee resistance	1949.5	0.662	0.04	n.s.
	Ethical concerns	1772.5	0.441	0.07	n.s.
	Customer distrust	1942.5	0.511	0.06	n.s.
Expected Business Impact	AI influence within three years	1242.5	0.001	0.37	Higher (younger)

Note: n.s. = not significant (p > 0.05).

summarizes the results of the Mann–Whitney U-tests for all 11 indicators.

Younger entrepreneurs reported significantly higher values across all indicators of perceived benefits, suggesting somewhat stronger confidence in AI’s potential to enhance productivity, innovation, and growth. The observed effects are small to moderate (r = 0.19–0.29), indicating modest but consistent differences rather than substantial behavioral gaps between age groups.

No significant differences were found in perceived risks. Both groups expressed comparable concern about data privacy, ethical dilemmas, and implementation challenges. This uniformity suggests that risk awareness is broadly shared, possibly reflecting common exposure to public discourse and policy narratives about AI.

Finally, younger entrepreneurs anticipate a significantly greater influence of AI on firm performance within the next three years (U = 1242.5, p < 0.001, r = 0.37), indicating a moderate-to-strong effect.

To assess whether the observed age-related differences persist after accounting for potential confounding factors, additional multivariate analyses were conducted using ordinal logistic regression models. This approach is appropriate given the ordinal nature of the dependent variables and allows the simultaneous inclusion of multiple control variables.

Two separate models were estimated. The first model (

Table 4. Ordinal logistic regression predicting Perceived Benefits of AI (controls included).

Predictor	B	SE	Wald χ2	p	OR (Exp(B))	95% CI for OR
Age group (18–34 = 1)	0.42	0.29	2.10	0.147	1.52	0.86–2.69
Gender (female = 1)	−0.18	0.26	0.48	0.488	0.84	0.50–1.40
Education (post-sec. or higher = 1)	0.51	0.24	4.47	0.034	1.67	1.04–2.68
Sector: Business services = 1	0.88	0.31	8.05	0.005	2.41	1.31–4.44
Sector: Consumer-oriented = 1	0.21	0.30	0.49	0.486	1.23	0.69–2.18

Note: Model fit: −2 Log Likelihood = 214.6; Model χ²(5) = 18.9, p = 0.002; Pseudo R² (Nagelkerke) = 0.18.

) examined perceived benefits of AI as the dependent variable, while the second model (

Table 5. Ordinal logistic regression predicting the expected business impact of AI (3 years).

Predictor	B	SE	Wald χ2	p	OR (Exp(B))	95% CI for OR
Age group (18–34 = 1)	0.97	0.32	9.17	0.002	2.64	1.41–4.95
Gender (female = 1)	−0.12	0.28	0.18	0.671	0.89	0.51–1.55
Education (post-sec. or higher = 1)	0.29	0.26	1.25	0.263	1.34	0.80–2.26
Sector: Business services = 1	0.39	0.33	1.40	0.237	1.48	0.78–2.81
Sector: Consumer-oriented = 1	0.05	0.31	0.03	0.874	1.05	0.57–1.93

Note: Model fit: −2 Log Likelihood = 198.3; Model χ²(5) = 22.6, p = 0.001; Pseudo R² (Nagelkerke) = 0.21.

) focused on the expected business impact of AI within the next three years. In both models, age group (18–34 vs. 35+), gender, education (post-secondary or higher), and sector affiliation were included as predictors. The sector was operationalized using dummy variables, with the reference category comprising extractive and transforming activities.

The overall model demonstrates good fit (Model χ²(5) = 18.9, p = 0.002; Nagelkerke R² = 0.18), indicating that the included predictors jointly explain meaningful variation in benefit perceptions.

When controlling for gender, education, and sector, the coefficient for age group remains positive but no longer reaches conventional levels of statistical significance (OR = 1.52, p = 0.147). This attenuation suggests that age-related differences observed in bivariate analyses are closely intertwined with entrepreneurs’ structural positioning. In particular, sector affiliation emerges as a strong predictor: entrepreneurs operating in business services are significantly more likely to perceive higher benefits of AI compared to the reference sector (OR = 2.41, p = 0.005). Educational attainment also shows a significant positive association with perceived benefits (OR = 1.67, p = 0.034).

Taken together, these findings indicate that differences in perceived benefits of AI operate partly through contextual and structural channels, notably industry domain and education, rather than reflecting a purely direct age-based effect.

The second model shows a strong overall fit (Model χ²(5) = 22.6, p = 0.001; Nagelkerke R² = 0.21). In contrast to the benefits model, age group remains a strong and statistically significant predictor of future impact expectations even after controlling for sector, education, and gender. Younger entrepreneurs (18–34) are substantially more likely to anticipate a positive impact of AI on business performance (OR = 2.64, p = 0.002). None of the control variables reaches statistical significance in this model. This pattern suggests that generational differences are particularly pronounced at the level of forward-looking strategic expectations, rather than in current evaluations of AI’s benefits.

Overall, the multivariate analyses refine the interpretation of the bivariate results. While age-related differences in perceived benefits of AI are partially accounted for by sectoral affiliation and educational attainment, age remains an independent and robust predictor of expectations regarding AI’s near-term business impact. These findings support the notion of an asymmetrical age gap in AI perceptions, whereby generational differences are more evident in future-oriented strategic outlooks than in contemporaneous benefit assessments.

5. Discussion and Implications

The findings provide clear answers to the study’s three research questions about entrepreneurs’ perceptions of AI: (1) younger entrepreneurs report higher perceived benefits of AI in bivariate analyses; (2) perceptions of AI-related risks show no significant age differences; and (3) expectations regarding AI’s future business impact are notably higher among younger founders. These patterns can be interpreted in light of the Slovenian entrepreneurial context, which is characterized by an advanced yet uneven digital ecosystem. According to the Digital Decade Country Report (European Commission, 2024), Slovenia demonstrates strengths in digital public services and connectivity infrastructure, including high e-Health maturity and fiber coverage compared to the EU average. At the same time, the digitalization of small and medium-sized enterprises and the adoption of advanced technologies such as AI and data analytics remain below the EU average, highlighting both opportunities and structural constraints in the national digital transition. This combination positions Slovenia as a digitally evolving context that is particularly relevant for examining how entrepreneurs perceive emerging technologies.

Although the national focus limits external validity, the identified relationships are likely to resonate with broader patterns observed in small, innovation-driven economies with comparable digital infrastructures. Together, these results confirm that entrepreneurs’ perceptions of AI differ by age, but this divergence is asymmetrical. Younger entrepreneurs express stronger opportunity-oriented evaluations of AI’s potential—reflected empirically in higher perceived benefits of AI in bivariate analyses, while risk perceptions remain remarkably stable across generations. This pattern—variation in opportunity-oriented evaluations alongside convergence in perceived risks—offers both theoretical and practical insights into how technology perceptions evolve within entrepreneurial contexts. Importantly, these results are consistent with the conceptual framework developed in the literature review (Section 2), which proposed that age influences entrepreneurs’ perceptions of AI primarily through differences in opportunity-oriented evaluations and motivational orientation, whereas risk perceptions are shaped by more universal social and informational influences.

Importantly, the inclusion of multivariate controls refines the interpretation of these age-related differences. While bivariate analyses indicate that younger entrepreneurs report higher perceived benefits of AI, the ordinal regression results show that this relationship is attenuated once sector affiliation and educational attainment are taken into account. This suggests that age-related differences in perceived benefits are closely intertwined with entrepreneurs’ structural positioning, particularly their concentration in AI-relevant sectors and differences in human capital. In this sense, age effects in benefit perceptions appear to operate partly through contextual channels rather than reflecting a purely direct age-based psychological disposition.

A contrasting pattern emerges for expectations regarding AI’s future business impact. Even after controlling for sector, education, and gender, age remains a strong and independent predictor of forward-looking expectations. Younger entrepreneurs are significantly more likely to anticipate a positive impact of AI on business performance within the next three years, indicating that generational differences are most pronounced at the level of strategic outlook and temporal orientation. Taken together, these findings point to an asymmetrical age gap in AI perceptions: current evaluations of benefits are shaped by structural context, whereas future-oriented expectations reflect more deeply rooted generational differences.

The findings also align with theoretical perspectives suggesting that opportunity-oriented evaluations and risk-related concerns may function as distinct dimensions of technology perception, rather than opposite ends of a single continuum. Although the present study cannot empirically test the dimensionality of these constructs due to the ordinal and single-item structure of the GEM indicators, the observed pattern of results is consistent with this conceptual distinction discussed in prior literature. While existing models such as TAM and TRI often conceptualize risk aversion as the inverse of perceived usefulness, the present findings suggest that entrepreneurs can simultaneously acknowledge AI-related risks while maintaining opportunity-oriented evaluations and positive future expectations regarding its business impact.

From a theoretical perspective, the results align with research suggesting that age-related differences in technology perception are not uniform, but depend on the specific evaluative dimension under consideration. Prior studies indicate that differences in motivational orientation, accumulated experience, and perceived time horizons shape how individuals interpret emerging technologies (Dwivedi et al., 2020; Kautonen et al., 2015; Levesque & Minniti, 2006; Parasuraman & Colby, 2015). In this study, these mechanisms appear particularly relevant for explaining age differences in future-oriented expectations regarding AI’s business impact. Younger entrepreneurs are more likely to adopt a forward-looking, growth-oriented perspective consistent with a promotion focus (Higgins, 1997), which predisposes them to anticipate the transformative effects of AI on their businesses.

By contrast, evaluations of AI’s current benefits appear more closely linked to entrepreneurs’ structural positioning and accumulated capabilities, rather than to age alone. Older entrepreneurs may exhibit a stronger prevention focus, emphasizing control, reliability, and the preservation of established systems; however, this orientation does not translate into systematically higher risk perceptions. The finding that both age groups evaluate AI-related risks similarly suggests that risk perception has become socially stabilized. Rather than being primarily age-driven, it increasingly reflects collective exposure to public discourse, regulatory frameworks, and professional norms surrounding AI (Dwivedi et al., 2020; Slovic, 2016).

These results broadly align with and extend technology adoption frameworks. In both the Technology Acceptance Model (TAM; Davis, 1989) and the Unified Theory of Acceptance and Use of Technology (UTAUT; Venkatesh et al., 2003, 2012), perceived usefulness and ease of use are typically stronger predictors of adoption intention than perceived risk. Consistent with this logic, the present findings show that generational differences in AI perceptions are not driven by divergent risk assessments but by differences in opportunity-oriented evaluations and, in particular, future-oriented expectations regarding AI’s business impact. While evaluations of current benefits are partly shaped by structural and contextual factors, younger entrepreneurs’ stronger expectations of AI’s future impact suggest greater adoption readiness at a strategic level.

Interestingly, this lack of age-based divergence in risk perceptions contrasts with earlier innovation-diffusion research (Morris & Venkatesh, 2000; Rogers, 2003), which often associated age with lower technological openness and greater uncertainty toward new technologies. The present results suggest that, in the case of AI, risk appraisal may be shaped less by age and more by the increasing institutionalization and mainstreaming of AI within digital ecosystems. As AI becomes embedded in public discourse, regulatory frameworks, and professional norms, exposure to shared information environments may lead to a convergence of risk perceptions across age groups.

The results further refine technology acceptance models (Davis, 1989; Venkatesh et al., 2003, 2012). Consistent with these frameworks, perceived usefulness—rather than perceived risk—emerges as the primary driver of adoption intention. In the present study, this pattern is reflected in the finding that age-related differences are not driven by divergent risk assessments, but by differences in opportunity-oriented evaluations and, more robustly, in future-oriented expectations regarding AI’s business impact. While evaluations of current benefits are partly shaped by structural and contextual factors, younger entrepreneurs’ stronger expectations of AI’s future impact point to greater adoption readiness at a strategic level.

At the same time, the similarity of risk perceptions across age cohorts suggests a decoupling of opportunity-oriented optimism and concern. Rather than functioning as opposite ends of a single continuum, these evaluative dimensions appear to coexist: entrepreneurs can acknowledge AI-related risks while remaining confident in its potential. This pattern aligns with emerging research on digital transformation, which increasingly recognizes the coexistence of caution and confidence in organizational responses to AI (Giuggioli & Pellegrini, 2023).

From a practical and policy perspective, the findings suggest that interventions focusing exclusively on raising awareness of AI-related risks may yield limited additional benefits, as such concerns appear to be widely shared across age groups. Instead, more differentiated capability-building initiatives are needed to address age-specific readiness gaps in AI adoption. For younger entrepreneurs, policy programs could emphasize risk literacy and responsible use, helping to balance strong future-oriented expectations with ethical and strategic foresight. For older entrepreneurs, experiential learning, peer-based mentoring, and hands-on demonstrations could strengthen digital self-efficacy and confidence in the operational relevance of AI.

Such targeted approaches acknowledge that generational differences are not primarily rooted in fear or risk aversion, but in differences in strategic outlook, motivational orientation, and perceived controllability. By addressing these distinct needs, policymakers and support institutions can reduce polarization between “AI enthusiasts” and “AI skeptics” and foster a more inclusive and effective innovation climate. At the ecosystem level, the findings highlight that successful AI diffusion in small and medium-sized enterprises (SMEs) depends not only on individual attitudes, but increasingly on the availability of enabling infrastructures, such as reliable data systems, interoperable digital tools, and affordable access to AI services (Iansiti & Lakhani, 2020; McKinsey & Company, 2023; OECD/BCG/INSEAD, 2025). As AI becomes embedded in organizational routines, entrepreneurs’perceptions of its usefulness appear to be shaped more strongly by contextual readiness and implementation conditions than by psychological disposition alone. This interpretation aligns with emerging research that conceptualizes AI adoption not merely as a technological act, but as a sociotechnical process shaped by learning systems, governance mechanisms, and entrepreneurial ecosystems (Verhoef et al., 2021; Vial, 2021). By situating individual perceptions within broader structural and institutional contexts, these perspectives help explain why age-related differences in AI perceptions may attenuate as enabling conditions improve.

Conceptually, the study provides empirical evidence supporting the view that opportunity-oriented evaluations of AI (theoretically conceptualized as optimism) and concern represent distinct yet coexisting dimensions of technology perception. The findings demonstrate that positive evaluations of AI’s potential do not imply lower risk awareness and that entrepreneurs can simultaneously acknowledge AI-related concerns while maintaining opportunity-oriented outlooks. By empirically distinguishing between current benefit evaluations, future-oriented expectations, and risk perceptions, the study refines existing models of entrepreneurial technology adoption and highlights the importance of treating these evaluative dimensions separately. This distinction offers a more nuanced perspective on digital readiness and suggests that generational differences in AI perception are best understood as asymmetrical and context-dependent rather than as uniform expressions of technological optimism or skepticism. Overall, this study contributes to the intersection of entrepreneurship, behavioral theory, and digital transformation by demonstrating that opportunity-oriented evaluations and concern toward AI function as distinct and coexisting cognitive dimensions, shaped by a combination of motivational orientation, experience, and institutional context. By empirically distinguishing between evaluations of current benefits, future-oriented expectations, and risk perceptions, the findings move beyond binary notions of “technological optimism versus skepticism.” Instead, they highlight how entrepreneurs interpret technological change through motivational and strategic lenses, and how generational differences in AI perception emerge in asymmetrical and context-dependent ways.

Several methodological limitations should be acknowledged. First, the study relies on cross-sectional survey data, which capture perceptions at a single point in time and therefore limit the ability to observe how entrepreneurs’ views on AI evolve with technological maturity and increased exposure. Longitudinal designs would allow future research to examine whether observed generational differences reflect stable life-course mechanisms, cohort effects, or shifts in the broader technological environment. Second, while the study combines bivariate nonparametric comparisons with multivariate ordinal regression analyses to account for key confounding factors such as sector affiliation, education, and gender, the scope for statistical control remains constrained by the relatively small sample size. This limits the inclusion of additional variables (e.g., entrepreneurial experience, firm age, digital capability) and precludes more complex modeling of mediation or interaction effects. Future studies with larger samples could employ more comprehensive multivariate frameworks, including structural equation modeling, to further disentangle the mechanisms underlying age-related differences in AI perceptions. Third, the measurement instruments used in the GEM survey rely on short Likert-type items that capture broad perceptions of AI benefits and risks. While appropriate for large-scale comparative research, these indicators provide limited granularity. Future research would benefit from more refined psychometric scales that capture nuanced aspects of AI literacy, risk appraisal, motivational orientation, and technological self-efficacy.

In addition, cross-country comparative studies could examine whether the asymmetrical patterns identified in this study depend on national digital infrastructures, cultural attitudes toward technology, or institutional trust in AI governance frameworks. Comparative designs would help assess whether the convergence in risk perceptions and the persistence of age-related differences in future-oriented expectations observed here are context-specific or generalizable across entrepreneurial ecosystems. Qualitative approaches, such as interviews or cognitive mapping, could further deepen understanding of how entrepreneurs interpret AI’s role in opportunity recognition, strategic decision-making, and business design. By addressing these methodological limitations and adopting more precise measurement instruments, future research can advance a more comprehensive and analytically robust understanding of how generational differences in entrepreneurial perceptions of AI emerge, evolve, and interact with contextual and institutional conditions.

A further methodological limitation concerns the measurement structure of the constructs. The GEM items used in this study were designed as broad ordinal indicators rather than as psychometrically validated multi-item scales. As such, they do not allow for a full confirmatory factor analysis (CFA) or an examination of latent correlations using polychoric matrices, which would provide stronger evidence of the dimensionality and distinctiveness of perceived benefits and perceived risks. Future research should therefore develop refined measurement instruments with multiple items per construct and greater response variability, allowing the use of advanced measurement models such as CFA, structural equation modeling, or estimators specifically suited for ordinal data (e.g., WLSMV). Such approaches would enable researchers to assess factor loadings, latent correlations, and measurement structure with greater precision and would provide a more rigorous test of whether benefit- and risk-related perceptions indeed reflect independent dimensions rather than opposite poles of a single continuum.

6. Conclusions

While based on Slovenian data, this study advances understanding of how age relates to entrepreneurs’ perceptions of AI by integrating technology adoption theory with insights from entrepreneurial psychology. Although the national focus limits external validity, the conceptual relationships identified are theoretically relevant across comparable entrepreneurial ecosystems. The findings highlight that age-related differences in AI perceptions are asymmetrical: evaluations of AI’s risks do not vary systematically by age, whereas differences emerge primarily in opportunity-oriented outlooks, particularly in expectations regarding AI’s future business impact. This pattern suggests that generational differences in how entrepreneurs interpret emerging technologies are driven less by fear of AI’s downsides and more by differences in strategic orientation, motivation, and perceived time horizons.

The study’s contribution lies in reframing age not as a determinant of technological conservatism, but as a contextual lens through which entrepreneurs interpret innovation. By linking opportunity-oriented evaluations, future-oriented expectations, and motivational orientation to perceptions of AI, the research extends theoretical understanding of how age shapes entrepreneurial sensemaking in the digital era, rather than adoption behavior per se.

Practically, these insights inform the design of AI-readiness and digital upskilling initiatives for entrepreneurs. Rather than adopting one-size-fits-all approaches, support programs should reflect age-related differences in strategic orientation and readiness. For younger entrepreneurs, interventions may place greater emphasis on risk awareness, ethical competence, and responsible use, complementing their strong future-oriented expectations. For older entrepreneurs, policies and support mechanisms should prioritize experiential learning, peer-based exchange, and accessible implementation tools that help translate awareness of AI into confident and practical application.

Future research should extend this inquiry through longitudinal designs to trace how entrepreneurs’ perceptions of AI evolve over time, as well as through comparative studies across national ecosystems to assess whether the asymmetrical patterns identified here reflect developmental, cohort-specific, or contextual effects. Such work would further illuminate how entrepreneurs’ cognitive and motivational, and contextual orientations shape the ongoing diffusion of AI across different economies.