Pedagogical Interaction and Social Values in Lifelong Learning in the Age of Artificial Intelligence

Abstract

The rapid integration of artificial intelligence (AI) accelerates the need for continuous skill acquisition. Consequently, this increases the importance of lifelong learning while raising fundamental questions about pedagogical interaction and human social values. To remain competitive, individuals must constantly acquire new skills and enhance existing ones. The aim of the article is to evaluate the stability of individual social value systems and the role of pedagogical interaction in lifelong learning during AI integration. The study uses a quantitative survey (N = 160) with a retrospective self-assessment model based on Schwartz’s Theory of Basic Human Values. The study processed data in IBM SPSS using non-parametric tests (Wilcoxon signed-rank, Kruskal–Wallis, Kendall’s rank correlation) to analyze how digital skills and sociodemographics influence technology perception. Findings reveal core value systems remain statistically stable; AI integration causes no internal value conflict. Digital skill level, rather than age, is the most significant factor in AI perception. While participants highly rate AI’s potential to customize learning, they express concerns regarding technological dependence. In the lifelong learning ecosystem, AI is viewed as a didactic tool rather than an educator replacement, as technology cannot provide essential social interaction and emotional support. Finally, higher education fosters a critical attitude toward AI’s ethical risks.

1. Introduction

Over the last decade, society has experienced significant scientific and technological breakthroughs, leading to an era of revolutionary change driven primarily by the rapid development of digital technologies, particularly artificial intelligence (AI). The accessibility of AI has fundamentally altered many aspects of human life, including the education sector (Gerlich, 2025). This progress has fostered a wide application of AI, ranging from everyday tools (Khlaif et al., 2023; Dai & Ke, 2022) to complex systems in specialized fields such as computer science and medicine (Bohr & Memarzadeh, 2020). Digital technologies are evolving rapidly. This process occurs alongside major geopolitical shifts. These combined factors impact the priorities and motivation of adult learners in Latvia. Previous research within the framework of the ‘IzVeTSKKEI’ project established that while core values remain resilient, external crises such as the COVID-19 pandemic and geopolitical tensions significantly heighten the individual’s focus on security (Balceraite & Lubkina, 2023; Balceraite et al., 2024). This study extends these findings by analyzing whether the same resilience applies to the current phase of rapid artificial intelligence integration. As a transformative technology capable of performing tasks previously considered achievable only by humans (Gruetzemacher & Whittlestone, 2022), AI integration into education (Kamalov et al., 2023; Seo et al., 2021) opens new opportunities for lifelong learning, promoting continuous skill adaptation and personalized learning (Hachem et al., 2026; Prayuda et al., 2026). However, the widespread proliferation of Generative AI (GenAI) creates a dual impact. On one hand, it offers substantial potential for efficiency and innovation (Bouhouita-Guermech et al., 2023; Feuerriegel et al., 2024; Elrawy & Wagdy, 2025). On the other hand, it poses serious ethical challenges, necessitating the careful management of new risks within the evolving educational landscape (Okoye et al., 2023; Mondragon-Estrada et al., 2023; Motlagh et al., 2023).

In today’s rapidly changing technological era, lifelong learning has become a global necessity that encompasses all aspects of life and supports continuous personal and professional development (Polat, 2025). Especially for employees engaging in lifelong learning, it is crucial to balance educational opportunities with professional commitments (Ahn, 2024). Workers must constantly acquire new skills and enhance existing ones to remain competitive. In the contemporary landscape, lifelong learning is no longer merely an economic instrument for maintaining competitiveness, but a fundamental right that ensures an individual’s social inclusion and personal growth throughout their lifespan (Mahoney & Kiernan, 2024). Consequently, this necessity creates a specific demand for innovative tools capable of simultaneously fostering both the learning process and knowledge construction. In this context, GenAI emerges as an invaluable resource. Its ability to support learning and improve knowledge creation offers innovative solutions (Shanmugasundaram & Tamilarasu, 2023) to mitigate the tension between professional duties and lifelong learning needs. Thus, GenAI facilitates effective skill development and adaptation to the evolving labor market, promoting continuous growth and competitiveness.

The impact of GenAI on society has become profoundly transformative, altering previous assumptions about it being a technology of the distant future. Today, GenAI has undeniably integrated into the activities of daily life, becoming an integral part of modern society (Neves & Almeida, 2024; Chen et al., 2023). This integration signifies a paradigm shift—GenAI is no longer merely an object of research but an active driver of socioeconomic processes.

Previous research by the authors, initiated in a conference report (Balceraite & Lubkina, 2023) and later expanded in a monograph (Balceraite et al., 2024), indicates that the individual social value system in the Latvian adult education space remains stable even during sharp external crises. This article continues the research series by analyzing value resilience in a new context—the integration of artificial intelligence.

The research objective is to evaluate the stability of an individual’s social value system and the role of pedagogical interaction in lifelong learning within the context of AI integration, analyzing how digital skills and sociodemographic factors influence technology perception and adaptation processes.

To conduct this study, the following research questions were formulated:

RQ1:

How stable is an individual’s core value system (based on the Schwartz Basic Values Model) under conditions of rapid AI integration?

RQ2:

How do an individual’s digital skills and sociodemographic factors influence the perception of AI-driven opportunities and risks within the learning process?

RQ3:

What is the significance of the human context and pedagogical interaction when evaluating the potential of AI to replace human resources in a digitalized learning environment?

2. Literature Review

In the dynamic era of technology, education has transformed into a continuous process that adapts to new challenges and opportunities. In recent years, the integration of computerized technologies and digital learning tools into the educational environment has significantly increased (Bentley et al., 2024). The pervasive presence of technology has fundamentally reshaped society, altering not only communication habits but also learning models themselves (Iivari et al., 2020; Awidi & Paynter, 2024). This digital transformation has reinforced the lifelong learning paradigm, providing individuals with the opportunity to acquire knowledge and skills in a flexible, continuous, and personalized manner (Timotheou et al., 2022).

As researchers point out (Ifenthaler et al., 2024), it is precisely the implementation of AI in educational processes in recent years that has initiated significant changes with the potential to substantially improve future learning outcomes. In scientific literature, the integration of AI, particularly within the context of higher education, is viewed as a promising approach capable of enriching study experience and fostering innovation, while effectively promoting problem-solving competencies (UNESCO, 2021).

2.1. The Concept and Potential of GenAI in Education

AI is a comprehensive concept that describes the ability of computer systems to simulate human cognitive functions, including perception, reasoning, and decision-making (Chiu et al., 2023; Resnik & Hosseini, 2024). AI tools are transforming the traditional lifelong learning environment by introducing innovative technological solutions. GenAI is a specific subfield of AI that, based on training with extensive sets of textual, audiovisual, and image data, is capable of autonomously generating new content (Watson et al., 2025). GenAI has become a dominant trend in educational research (Montenegro-Rueda et al., 2023; Giannakos et al., 2024); it is capable of mimicking the learning process and offering independent problem-solving solutions.

Scientific research (Zawacki-Richter et al., 2019) predicts that the rapid evolution of GenAI technologies will facilitate their deeper integration into education. This is confirmed by statistical data: in 2024, the global number of GenAI users reached 314.38 million, with a projected increase to 729.11 million by 2030 (Statista, 2024). Significantly, a 2024 international student survey revealed that 86% of respondents already use GenAI tools in the study process (Statista, 2025), which confirms the widespread adoption of these technologies. The most popular tools include chatbots, content and image generators, as well as language correction tools (Quinn & Burns, 2025), the demand for which is driven by both their efficiency and accessibility (including free-of-charge versions).

Studies reveal that the integration of GenAI with high-quality didactic materials can significantly improve the efficiency of the learning process (Tzirides et al., 2024). As noted by Adiguzel et al. (2023), technologies such as adaptive learning systems and ChatGPT (GPT-3.5) provide personalized learning strategies that promote academic achievement. Tlili et al. (2023) emphasize the role of ChatGPT in improving learning outcomes by providing structured information in an easily accessible manner, which fosters deeper understanding and marks a paradigm shift in lifelong learning. Researchers (Ifenthaler et al., 2024; Tlili et al., 2023; Adiguzel et al., 2023) acknowledge the potential of GenAI in expanding educational accessibility and effectively identifying knowledge gaps.

At the same time, research emphasizes that public digital literacy and awareness are prerequisites for the responsible implementation of GenAI (Southworth et al., 2023; Stolpe & Hallström, 2024). Issues related to the objectivity, reliability, and ethical use of information are becoming increasingly relevant. The public availability of ChatGPT has sparked discussions regarding the revision of traditional pedagogical teaching methods (Tlili et al., 2023). Although institutions face cybersecurity and copyright risks (Khlaif et al., 2023), the potential of technology is still considered to be high. The consensus among current researchers (Adiguzel et al., 2023; Tlili et al., 2023) suggests that GenAI is shifting from a mere technical aid to a personalized learning architect. However, a critical gap remains in understanding whether this personalization compromises the learner’s intrinsic motivation, a role traditionally fulfilled by the human educator.

2.2. The Impact of Artificial Intelligence on Lifelong Learning

The modern labor market is being fundamentally transformed by the rapid integration of AI and automation. World Economic Forum data indicate that in the coming decade, approximately half of all workers will require reskilling to maintain their competitiveness. These technological changes create an imperative demand for high-level digital skills, which are a prerequisite for the effective use of AI tools (Vesnic-Alujevic & Saitis, 2025). It is projected that in the future, digital competencies will be necessary for 90% of all employees. Already, a gap is identifiable between the existing skills of the workforce and market demands, particularly in the AI sector, making workforce reskilling a critical priority. In this context, active discussions are taking place within the academic community regarding whether the traditional “digital divide” is still to be considered a generational issue, or if, under the influence of digitalization, it has transformed primarily into a matter of competencies and skills (Bentley et al., 2024; Piwowar-Sulej et al., 2024). Within the context of lifelong learning, senior education is equally significant, as it promotes cognitive resilience, social inclusion, and the maintenance of quality of life (Findsen & Formosa, 2021).

The integration of AI into the lifelong learning ecosystem creates both innovative opportunities and challenges (Adiguzel et al., 2023; Mouta et al., 2024). GenAI is increasingly being integrated into pedagogical processes (Pisica et al., 2023), transforming learning approaches through personalization (Kamalov et al., 2023; Al-Zahrani & Alasmari, 2024) and promoting interactive educational accessibility. Personalized and interactive learning methods are essential in lifelong learning. These approaches maintain motivation and support individual growth, regardless of the learner’s prior experience. Furthermore, AI technologies offer adaptive solutions that facilitate learning for older adults as well.

Emphasizing the importance of lifelong learning, researchers (Stolpe & Hallström, 2024; Hachem et al., 2026) put forward two recommendations for the continuous improvement of societal AI literacy: first, the need to maintain an understanding of the operating principles of GenAI and its changing social impact; and second, the necessity of fostering career adaptation for adults in AI-related sectors. This need for learning throughout life is further supported by the insights of researchers (Southworth et al., 2023; Polat, 2025; Hachem et al., 2026) regarding the value of experiential learning, for example, by providing opportunities for workers and other adults to apply GenAI in solving real-life and workplace problems.

Despite technological progress, research (Seo et al., 2021; Laupichler et al., 2022) points toward certain limitations. While AI streamlines online learning, it cannot fully replace human interaction, which is a critical element in value education and lifelong learning, as well as in the development of critical thinking and social skills (Han et al., 2022). For example, in a study by Denecke et al. (2023), although 80% of respondents confirmed using AI tools, the results indicate that this technology is still in the adaptation phase, and its full potential in the academic environment remains untapped. Furthermore, concerns exist in the literature regarding the ability of technology to sustain a learner’s intrinsic motivation, which has traditionally been the role of the educator (Paloș et al., 2025; Seo et al., 2021).

2.3. Transformation of Individual Values in Lifelong Learning

The beginning of the 21st century is characterized by dynamic technological, economic, and social transformations, in which AI is becoming the central driving force (Alemayehu Tegegn, 2024). Technological progress and global digitalization are fundamentally changing individual value systems and worldviews, presenting new challenges for pedagogical interaction within the context of lifelong learning. The pluralism of adult values promotes societal tolerance and cultural expansion. However, value bubbles created by AI algorithms can hinder social consensus on fundamental issues.

Values as abstract beliefs that influence human action and choice have been studied since antiquity; however, in the present day, facing automation and algorithmic decision-making, scholarly interest in them has increased significantly (Hanel et al., 2018). Values are a complex, interdisciplinary concept encompassing both individual and social aspects that directly affect an individual’s actions, decisions, and development in the digital environment (Ponizovskiy et al., 2019; Toker Gökçe, 2021). They reflect the moral and ethical principles that serve as the foundation for human behavior. Understanding these principles is essential for building a coherent society and ensuring sustainable development in an era where technology is taking over a portion of human functions (Bednar & Spiekermann, 2024; Schwartz, 2012).

In scientific literature, the model developed by Shalom Schwartz is frequently employed for the analysis of values, defining them as motivational goals that serve as guiding principles in an individual’s life. Although values are traditionally considered to be relatively stable beliefs, their transformation may become more rapid within the modern educational space where AI is integrated. As noted by Bardi (2011), targeted educational programs and technological tools can serve as instruments for replacing outdated beliefs with new values appropriate for the era. This creates a necessity for a new form of pedagogical interaction, in which the educator and technology together create an environment for the critical re-evaluation of values.

Dzalbe et al. (2015) emphasize the balance between value stability and changeability: if values were absolutely static, social development would not be possible; conversely, excessive changeability would create adaptation difficulties. This balance becomes particularly fragile under the influence of AI, where values serve as an instrument for the development of an individual’s skills and competencies, allowing for a more effective response to upcoming technological shocks (Bojanowska & Urbańska, 2021). Schwartz (2012) adds that values motivate action, which can have both positive and negative consequences, depending on how the individual interacts with the social and technological environment.

In the pedagogical process, a significant role is played by society’s attitude toward value groups, as an unclear understanding can lead to an individual’s value disorientation (Gupta, 2016). Research indicates a correlation between education level and value priorities: for individuals with higher education, values such as security and conformity are less pronounced, whereas self-actualization and achievement dominate (Muzikante, 2011). Building upon the ‘IzVeTSKKEI’ project’s findings regarding value resilience during pandemic and geopolitical crises, this study positions AI integration as a unique form of technological transformation. It analyzes whether core values serve as the same psychological anchor in the face of algorithmic shifts as they did during physical security threats.

3. Materials and Methods

3.1. Research Method

Within the framework of the study, a survey was used as the data collection method, which allowed for the acquisition of quantitative data regarding respondents’ subjective self-assessment and attitudes toward the role of AI in lifelong learning. The obtained data provide insight into the respondents’ experiences and are essential for analyzing the cognitive and behavioral patterns that determine the acquisition and integration of new technologies, including AI, into professional development.

Empirical data collection was conducted from October 2024 to February 2025 using an online instrument created on the “Google Forms” platform. To ensure a more representative sample, the survey was distributed via social media environments. To achieve the research objective, a questionnaire was constructed and structured into two parts: (1) profile questions; (2) an assessment of the impact of AI based on ethical and social prerequisites (for the detailed survey structure and methodological framework, see Appendix A). To measure value priorities, respondents were asked to rank their values on a 10-point rating scale, where the ranking reflected the relative importance of each dimension. The methodological framework, grounded in S. Schwartz’s value model, builds upon a previously tested survey instrument used to assess human values under the influence of external conditions in Latvia (Balceraite & Lubkina, 2023). The study employed a retrospective self-assessment model, in which respondents were asked to evaluate their value priorities at two points of contact: the current state (V2) and a retrospective self-assessment of the state prior to the broad integration of AI tools into daily life (V1).

This approach allows for a consistent comparison with prior research stages, capturing the subjectively perceived value dynamics and an individual’s conscious sense of stability during different periods of social and technological transformation. It is a vital factor in the process of psychological adaptation. The chosen retrospective model does not provide an objective measurement of long-term value sequences. Instead, it aims to capture value stability as a subjectively perceived defense mechanism against technological stress.

Data analysis included an evaluation of the respondents’ education level, computer skills, and attitudes toward AI. To ensure methodological rigor, the internal consistency of the survey instrument was assessed by calculating Cronbach’s alpha for each distinct thematic subscale independently, rather than for the questionnaire as a whole. The retrospective self-assessment scale of individual values (V1) and the current state assessment scale (V2) demonstrated excellent reliability (α = 0.910 and α = 0.926, respectively). Furthermore, the subscales evaluating AI’s potential benefits and potential challenges in the learning process showed high internal consistency (α = 0.876 and α = 0.873, respectively). Finally, the subscales for general AI impact assessment and the significance of social interaction yielded reliability coefficients of α = 0.866 and α = 0.851, confirming the robust reliability of the theoretical constructs measured in this study. Since the results of the Kolmogorov–Smirnov test showed a significant deviation from a normal distribution (p < 0.05), indicating that the results do not follow a normal distribution, non-parametric tests will be used in the inferential statistics.

Statistical data processing was performed using IBM SPSS Statistics 31.0 software. The study employed descriptive statistics, as well as the Kruskal–Wallis test to determine differences and Kendall’s rank correlation for relationship analysis. Additionally, the Wilcoxon signed-rank test was used to evaluate value stability within the dynamics.

3.2. Research Sample

The study sample was developed by ensuring respondent diversity according to three primary criteria: (1) socioeconomic status (including salaried employee, entrepreneur, student, homemaker, unemployed, retiree), (2) attained educational level, and (3) geographical coverage across all regions of Latvia.

A total of 160 respondents participated in the study (N = 160). Analyzing the geographical distribution, the majority of respondents were from Riga (n = 63; 39.4%), followed by the Riga region and Latgale (n = 23; 14.4% each), Kurzeme (n = 22; 13.8%), Vidzeme (n = 16; 10%), and Zemgale (n = 13; 8.1%). In the gender structure, women predominated (n = 144; 89.6%), while the proportion of men was 10.4% (n = 16).

The age of the respondents ranged from 25 to 72 years, with a mean of 43.35 years (M = 43.35, SD = 11.92). The analysis of the age structure reveals the following distribution: 25–30 years (n = 27; 16.9%), 31–35 years (n = 20; 12.5%), 36–40 years (n = 29; 17.1%), 41–45 years (n = 20; 12.5%), 46–50 years (n = 21; 13.1%), 51–55 years (n = 14; 8.8%), 56–60 years (n = 14; 8.8%), 61–65 years (n = 10; 6.3%), and 66+ years (n = 5; 3.1%). In terms of educational attainment, the majority of respondents held a higher education degree: a Master’s degree (n = 59; 36.9%) and a Bachelor’s degree (n = 50; 31.3%). First-level higher education was held by n = 22 (13.8%), a Doctoral degree by 2.5% (n = 4), while 15.6% (n = 25) of respondents had not attained a higher education degree.

The distribution of respondents according to their self-assessed level of digital skills is summarized in

Table 1. Respondent Profile by Level of Computer Proficiency ¹.

Level	n (%)
Basic level (Ability to work independently with a computer, use text editors (Word), spreadsheets (Excel), presentation programs (PowerPoint), search for information on the Internet, use e-mail and basic file management functions)	42 (26.3%)
Intermediate level (In addition to basic skills, the ability to use more complex programs and functions, such as creating formulas in Excel, creating more complex presentations, using databases, analyzing data, and working with different file formats)	97 (60.6%)
High level (In-depth knowledge of computer programs and operating systems, ability to automate processes, create complex analyses, program, work with specialized programs (e.g., statistical programs, CAD programs)	21 (13.1%)

¹ Adapted from the European Digital Competence Framework ‘DigComp 2.2’ (Vuorikari et al., 2022).

.

Overall, the demographic data of the sample indicate a high educational level among respondents, which correlates with their digital skills. The majority of respondents (68.2%) have attained at least a Bachelor’s or Master’s degree, suggesting a qualified and academically educated target audience.

Respondents with higher education or intermediate digital skills constitute the group with the highest potential for the effective integration of AI tools into lifelong learning processes. Prior experience with complex software creates favorable conditions for acquiring AI skills. This background facilitates the use of personalized learning platforms and specialized information retrieval. At the same time, it should be noted that despite the overall digital competence of the sample, the portion of respondents with basic skills (26.3%) may face technological barriers and greater challenges in integrating AI into their professional development.

The study was conducted in accordance with ethical standards pertaining to research involving human subjects. Participation in the survey was voluntary, and before completing the questionnaire, respondents were informed of the research objective, the use of data, as well as their right to withdraw from participation at any time without any negative consequences.

To ensure full anonymity and confidentiality, no personally identifiable information was collected; specifically, respondents were not required to provide email addresses or names. Access to the data was restricted, ensuring it was available only to the lead researcher responsible for data collection and SPSS processing. The obtained data were stored securely and used solely for the specified academic purposes.

4. Results

4.1. Transformation of Individual Social Values Driven by AI

Within the framework of the study, an assessment of individual values was conducted in two stages: prior to the widespread use of AI (V₁) and after its implementation (V₂). The results, summarized in

Table 2. Comparison of Individual Value Indicators Before and After the Widespread Integration of AI (N = 160).

Value	Before (V1) M (SD)	After (V2) M (SD)	Before (V1) Mdn (IQR)	After (V2) Mdn (IQR)	z	p
Security	8.07 (2.48)	8.12 (2.45)	9.00 (3.00)	9.00 (3.00)	−0.60	0.548
Conformity	6.98 (2.48)	7.14 (2.50)	8.00 (4.00)	8.00 (4.00)	−1.37	0.172
Tradition	7.07 (2.48)	7.11 (2.59)	8.00 (4.00)	8.00 (3.00)	−0.10	0.917
Benevolence	7.81 (2.26)	7.76 (2.37)	8.00 (3.00)	8.00 (3.00)	−0.62	0.534
Universalism	7.11 (2.25)	7.16 (2.30)	7.05 (3.00)	8.00 (3.00)	−0.39	0.695
Self-Direction	7.71 (2.28)	7.61 (2.32)	8.00 (4.00)	8.00 (4.00)	−0.89	0.375
Stimulation	6.81 (2.49)	6.93 (2.38)	7.00 (4.00)	7.00 (4.00)	−0.34	0.736
Hedonism	6.21 (2.60)	6.32 (2.69)	7.00 (4.00)	7.00 (4.00)	−0.55	0.580
Achievement	6.98 (2.51)	7.00 (2.68)	7.00 (4.00)	8.00 (4.00)	−0.18	0.861
Power	4.75 (2.75)	5.04 (2.85)	5.00 (5.00)	5.00 (5.00)	−1.59	0.113

V₁—value assessment before AI; V₂—value assessment after AI; M—arithmetic mean; SD—standard deviation; Mdn—median; IQR—interquartile range; z—Wilcoxon test statistic; p—statistical significance.

, characterize the distribution of value priorities and their stability or variability under the influence of technological transformation.

The analysis of both measurements using the non-parametric Wilcoxon signed-rank test revealed that the respondents’ value system remains relatively stable. As shown in Table 2, all p-values exceed the threshold of 0.05, confirming the stability of the core value system. Although a specific shift in indicators can be observed, the results show that for none of the ten individual values studied were there any statistically significant differences (p > 0.05).

In both stages, the highest-rated value is “Security” (M_V1 = 8.07, SD = 2.48; M_V2 = 8.12, SD = 2.45), where the median score (Mdn = 9.00, IQR = 3.00) indicates a pronounced data shift toward higher ratings. Similarly, high priority is assigned to “Benevolence” (M_V1= 7.81, SD = 2.26; Mdn = 8.00) and “Self-Direction” (M_V1 = 7.71, SD = 2.28; Mdn = 8.00).

In contrast, the lowest ratings were recorded for the value of “Power” (M_V1 = 4.75, SD = 2.75; M_V2 = 5.04, SD = 2.85). Although this value showed the largest increase in the arithmetic mean, the large interquartile range (IQR = 5.00) indicates high variability of opinions among respondents, and the Wilcoxon test confirms that this increase is not statistically significant (z = −1.59; p = 0.113). While minimal fluctuations in means and medians are observed in certain dimensions, they do not reach the threshold of statistical significance (p > 0.05), which confirms the stability of core values.

The lack of statistically significant differences (p > 0.05) between the V1 and V2 measurements indicates that respondents perceive their value system as stable. Consequently, the emergence of AI has not created an internal value conflict in their consciousness. This suggests high internal integrity in the context of technological change. Even such a rapid technological transformation as that created by AI does not cause drastic changes in the value hierarchy, confirming the previously observed trend of value system resilience. Furthermore, it indicates that core values within the framework of lifelong learning remain resilient even under conditions of rapid technological transformation.

Lifelong learning in the digital age is closely linked to individual adaptation, where both age and digital competence play a critical role. To evaluate respondents’ attitudes toward the integration of AI specifically within the learning environment, the study identified two essential sets of criteria: the potential of AI in a personalized learning process (adaptation of materials and pace, skill diagnostics, and promotion of motivation) and potential challenges (development of dependency, risks of inequality, and the reduction in social interaction).

4.2. The Role of Digital Skills and Demographics in the Perception of AI in the Learning Process

To evaluate perceptions of AI-driven changes, the study analyzed its impact on six essential aspects of daily life and society using descriptive statistics and Kruskal–Wallis group comparisons (see

Table 3. Respondents’ Assessment of the Impact of AI in Various Fields: Means and Standard Deviations (N = 160).

Criteria	Descriptive Statistics				Statistical Significance (p-Value)
	M	SD	Mdn	Mo	Age	Education	Computer Skills
Information Flow	3.84	1.08	4.00	4.00	0.226	0.063	0.004
Labor Market Automation	3.39	1.10	4.00	4.00	0.730	0.708	0.045
Personal Adaptation	3.24	1.00	3.00	3.00	0.768	0.519	0.003
Social Inequality	2.71	1.08	3.00	3.00	0.943	0.170	0.029
Ethical Aspects	3.08	1.14	3.00	4.00	0.993	0.033	0.045
Educational Environment	3.78	1.10	4.00	4.00	0.705	0.719	<0.001

). The complete set of p-values is reported to ensure statistical transparency and provide empirical evidence for the observed stability of opinions across age groups and educational backgrounds, even where differences did not reach statistical significance (p > 0.05).

Analyzing the respondents’ opinions on the impact of AI in various fields, it is observed that, overall, the role of AI in information accessibility and retrieval was rated highest (M = 3.84, SD = 1.08), indicating that the majority of respondents tend to agree that AI significantly improves information flow. A relatively more cautious view is observed regarding the intensification of social inequality (M = 2.71, SD = 1.08), where the median and mode (Mdn = 3.00, Mo = 3.00) suggest a neutral or expectant stance. It should be noted that the standard deviation is relatively high (above 1.00) across all criteria, indicating a diversity of opinions among respondents—there is no consensus even on the unambiguity of AI’s impact.

To determine whether statistically significant differences exist in respondents’ opinions based on age, education, and computer skills, the Kruskal–Wallis test was employed. The data indicate that respondents’ age is not a decisive factor in their attitude toward AI; no statistically significant differences were found in any of the analyzed criteria (p > 0.05). This leads to the conclusion that both younger respondents and seniors perceive the impact of AI similarly.

In contrast, the level of education showed a statistically significant difference in only one aspect—ethical issues (p = 0.033). This suggests that the level of education influences the degree of importance respondents assign to ethical challenges in the context of AI. In other aspects, including labor market issues and information retrieval, education level had no statistically significant impact.

The most significant finding of the study is the close link between respondents’ computer skills and their opinions. The analysis revealed statistically significant differences across all analyzed criteria (p = 0.05). A particularly pronounced difference is observed in the educational environment aspect, where the level of significance is highest (p = 0.001). Differences regarding information accessibility (p = 0.004) and personal adaptation (p = 0.003) are also highly compelling.

Using the criteria from Table 3, the conducted Kendall’s tau-b correlation analysis revealed statistically significant positive correlations across all analyzed aspects of AI impact (p < 0.001). The correlation coefficients range from τ = 0.229 to τ = 0.520, indicating weak to moderate associations between the variables.

The strongest correlation (τ = 0.520) is observed between the impact of AI on labor market automation and personal adaptation. This suggests that respondents perceive these processes as interrelated—the more significant the perceived impact of automation, the higher the perceived necessity for personal adaptation. Similarly, a high correlation (τ = 0.511) exists between information accessibility and labor market changes. The results indicate consistency in respondents’ views. For example, those who perceive a significant impact of AI in information retrieval tend to perceive similar impacts in areas such as education and ethics.

To determine the causes of these opinions and identify the decisive factors, an in-depth correlation analysis was conducted in the further stage of the study, including independent variables: respondents’ age, level of education, and computer skills. This analysis will allow for an examination of whether, and to what extent, these demographic and competence indicators correlate with the respondents’ assessment of AI’s impact. The results of Kendall’s correlation analysis between demographic indicators and AI assessment are presented in

Table 4. Kendall’s Correlation Coefficients Between Demographic Indicators and AI Assessment.

Criteria	$\mathbf{Age},\mathit{\tau }$	$\mathbf{Education},\mathit{\tau }$	$\mathbf{Computer}\mathbf{Skills},\mathit{\tau }$
Information Flow	−0.126 *	0.031	0.203 **
Labor Market Automation	−0.089	0.010	0.134
Personal Adaptation	−0.056	−0.045	0.184 **
Social Inequality	0.029	0.097	0.086
Ethical Aspects	0.037	0.203 **	0.098
Educational Environment	−0.089	0.036	0.052

* Correlation is significant at the 0.05 level (2-tailed, p < 0.01). ** Correlation is significant at the 0.01 level (2-tailed, p < 0.01).

.

Kendall’s tau-b correlation analysis identified a statistically significant positive correlation between respondents’ computer skills and two aspects of AI: information accessibility (τ = 0.203; p = 0.004) and personal adaptation (τ = 0.184; p = 0.009). The level of education also showed a statistically significant positive correlation with the perception of ethical issues (τ = 0.203; p = 0.002). Regarding demographic factors, a weak negative correlation was identified between age and the assessment of information accessibility (τ = −0.126; p = 0.038), suggesting that older respondents tend to rate this aspect lower.

A detailed cross-tabulation analysis (see Information Flow) reveals that this trend is not driven by opposition to AI, but by a shift in intensity. While 73.3% of respondent’s overall view AI’s impact on information flow positively, a significant deviation was identified in the 51–55 age group (Adjusted Residual = 2.3), where 36.4% of respondents adopted a neutral stance. This suggests that while older generations do not directly oppose AI, they exhibit a higher degree of caution or uncertainty regarding its practical benefits in information management.

Table 5. Respondents’ Assessment of AI Potential and Challenges in the Learning Process (N = 160).

Criteria	Descriptive Statistics				Statistical Significance (p-Value)
	M	SD	Mdn	Mo	Age	Education	Computer Skills
Potential Benefits
Content Adaptation	3.76	1.15	4.00	5.00	0.845	0.238	0.067
Pace Individualization	3.72	1.14	4.00	4.00	0.908	0.487	0.133
Skill Diagnostics	3.72	1.17	4.00	5.00	0.790	0.345	0.480
Motivation Enhancement	3.26	1.27	3.00	3.00	0.338	0.662	0.112
Potential Challenges
Technological Dependency	4.21	1.03	5.00	5.00	0.392	0.793	0.660
Risk of Inequality	3.63	1.17	4.00	3.00	0.523	0.886	0.722
Threats to Social Interaction	3.87	1.07	4.00	5.00	0.612	0.672	0.844

summarizes these results, illustrating how digital skills and demographic factors influence the respondents’ general assessment of AI.

Analyzing the respondents’ attitudes toward the integration of AI specifically within the learning environment (see Table 5), a tendency is observed where respondents perceive high risks alongside potential benefits.

Among the opportunities provided by AI, respondents rated content adaptation to individual needs (M = 3.72, Mo = 5.00) and the individualization of learning pace (M = 3.72) highest. The mode (5.00) indicates that a large proportion of respondents “strongly agree” that AI can help personalize content. In contrast, the most cautious assessment was observed regarding AI’s ability to enhance learner motivation (M = 3.26, Mo = 3.00). The general indicator from the mode suggests a neutral stance, showing that respondents doubt whether technology can address the problem of lack of motivation.

Crucially, the highest mean score in the entire survey was received not by a benefit, but by the risk of technological dependency (M = 4.21, SD = 1.03). Both the mode (5.00) and the high mean clearly demonstrate that this is the respondents’ primary concern. Similarly, the threat to social interaction (M = 3.87) is perceived as a significant risk, while the risk of inequality was rated lower (M = 3.63).

In contrast to the general assessment of AI’s impact, no statistically significant differences between demographic groups were found regarding specific questions on the learning process. The Kruskal–Wallis test indicated that neither age, education, nor computer skills significantly influence respondents’ opinions (p > 0.05 across all criteria). For instance, concerns about technological dependency are equally high among respondents with both low and high computer skills (p = 0.660), and the potential to adapt learning materials is rated similarly by representatives of all age groups (p = 0.845).

The correlation analysis based on the criteria in Table 3 reveals two significant trends in respondents’ opinions regarding the role of AI in the learning process. First, a high degree of internal consistency is observed within both thematic blocks. In the potential benefits block, all items correlate with each other with high statistical significance (p < 0.001) and strong coefficients (up to τ = 0.705). This indicates that respondents who perceive AI potential in one aspect (e.g., content adaptation) consistently perceive it in others as well (motivation, pace). An identical situation is observed in the challenges block, where risks correlate closely with each other (e.g., τ = 0.655 between inequality and social interaction). Second, no statistically significant correlation was found between the perception of benefits and the perception of risks.

To determine what exactly shapes this dual perception, the study will continue with an in-depth correlation analysis, incorporating external factors. An examination will be conducted on how respondents’ age, level of education, and computer skills correlate with each of these blocks separately, in order to identify which demographic indicators foster optimism (focusing on benefits) and which foster caution (focusing on risks).

The correlation analysis (see

Table 6. Correlations Between Demographic Factors and Perceived Impact of AI on the Learning Process.

Criteria	$\mathbf{Age},\mathit{\tau }$	$\mathbf{Education},\mathit{\tau }$	$\mathbf{Computer}\mathbf{Skills},\mathit{\tau }$
Potential Benefits
Content Adaptation	−0.082	−0.054	0.082
Pace Individualization	0.030	0.039	0.099
Skill Diagnostics	0.006	0.002	0.076
Motivation Enhancement	−0.023	0.011	0.091
Potential Challenges
Technological Dependency	0.057	0.047	−0.005
Risk of Inequality	0.053	0.068	0.020
Threats to Social Interaction	0.073	0.074	0.040

) revealed that regarding questions on AI application in the learning process and its associated risks, no statistically significant relationship exists with respondents’ demographic indicators. None of the calculated correlations are statistically significant (p > 0.05), and the coefficient values are very low (ranging from −0.082 to 0.171). This leads to the conclusion that respondents’ opinions on AI in education are universal within this sample.

Neither age, level of education, nor computer skills determine how optimistically or cautiously respondents view AI in learning. It is crucial to emphasize that, in contrast to the general perception of AI (where the influence of skills was previously observed), computer skills are no longer a decisive factor specifically in the educational context. Both technologically proficient respondents and beginners equally perceive the potential of AI to adapt learning materials and share equal concerns regarding the risk of technological dependency.

4.3. The Impact of AI on the Role of Pedagogical Interaction

To determine whether, in the era of AI and digital tools, lifelong learning participants still value direct contact with instructors and other participants in the learning process, an analysis of the significance of interaction was conducted.

Table 7. Respondents’ Assessment of the Significance of Social Interaction in the Learning Process and the Influence of Demographic Factors (N = 160).

Criteria	Descriptive Statistics				Statistical Significance (p-Value)
	M	SD	Mdn	Mo	Age	Education	Computer Skills
Importance of Feedback	4.38	0.93	5.00	5.00	0.737	0.183	0.148
Collaboration and Joint Problem-Solving	4.02	1.11	4.00	5.00	0.840	0.399	0.570
Creating a Positive Learning Environment	4.26	1.00	5.00	5.00	0.166	0.544	0.278
Learner Autonomy (Independence from Contact)	3.03	1.27	3.00	2.00	0.066	0.369	0.433
Potential of Technology to Replace Human Resources	2.40	1.41	2.00	1.00	0.272	0.576	0.938

summarizes the descriptive statistics on the role of communication and provides a group difference analysis to examine whether respondents’ age, education, and digital skills influence their opinions on the ability of technology to replace human resources in adult education.

Analyzing the respondents’ opinions on the role of human contact in the learning process, the obtained data outline a clear socio-centric paradigm that dominates regardless of the respondents’ digital competence or age. Lifelong learning participants assign critically high importance specifically to interaction with instructors and peers. The highest ratings were given to receiving feedback (M = 4.38, Mo = 5.00) and creating a positive learning environment (M = 4.26, Mo = 5.00). This indicates that the learning process is not perceived merely as information consumption (which could be provided by AI), but rather as a social process in which emotional support and quality assessment from the educator play a decisive role.

Although overall there are no statistically significant differences (p > 0.05), the mean ranks reveal specific trends. In the active working age group (36–40), the need for feedback dominates (Mean Rank = 87.66), while for seniors (66–72), collaboration with peers is more important (Mean Rank = 95.10), serving a socialization function. The most pronounced differences (p = 0.066) are observed in autonomy: the youngest group (31–36) tends toward independence (Mean Rank = 107.37), whereas pre-retirement age respondents (51–56) prefer a guided process (Mean Rank = 66.50).

When evaluating the ability to learn autonomously without contact (M = 3.03), respondents took a neutral position with a tendency toward disagreement (Mo = 2.00). This suggests that although adult education participants are capable of independent learning, they do not consider it the most effective or desirable method.

The most significant finding of the study in this section is the absence of statistically significant differences (p > 0.05) across all demographic groups. Both younger respondents and seniors value contact equally high (p-values range from 0.066 to 0.840). Even respondents with high computer skills (p > 0.148) do not believe that technology can replace humans. Furthermore, the level of education does not change this view (p > 0.183).

To gain an in-depth understanding of the internal relationships between various aspects of communication in the learning process, Kendall’s correlation analysis was conducted based on the previously defined criteria (see

Table 8. Relationship Between Demographic Variables and Respondents’ Evaluation of Social Interaction in Learning.

Criteria	$\mathbf{Age},\mathit{\tau }$	$\mathbf{Education},\mathit{\tau }$	$\mathbf{Computer}\mathbf{Skills},\mathit{\tau }$
Importance of Feedback	−0.059	0.026	0.023
Collaboration and Joint Problem-Solving	0.095	0.114	−0.052
Creating a Positive Learning Environment	0.045	0.083	−0.097
Learner Autonomy (Independence from Contact)	−0.156 **	−0.130 *	0.082
Potential of Technology to Replace Human Resources	−0.007	−0.075	0.000

** Correlation is significant at the 0.01 level (2-tailed) (p < 0.01); * Correlation is significant at the 0.05 level (2-tailed) (p < 0.05).

). The analysis allows for the identification of how respondents’ beliefs regarding the importance of communication and the role of technology are interrelated.

The data reveal two distinct, separate clusters of factors. The first three indicators—feedback, collaboration, and a positive environment—form a very strong, interconnected block. A strong, statistically significant positive correlation exists between these variables (τ ranges from 0.570 to 0.641, p < 0.001). This implies that respondents who value feedback from an educator almost invariably also value collaboration with peers and the emotional atmosphere. These aspects are inseparable in the respondents’ perception—for them, quality learning is inherently a social process.

A strong positive correlation (τ = 0.485, p < 0.001) was found between the ability to learn autonomously and the belief that technology can replace humans. This indicates a clear trend: the more confident a respondent is in their ability to learn independently (without contact), the more open they are to the idea of a fully automated, AI-driven learning process. In this context, autonomy serves as a “gateway” to instructor replacement.

Furthermore, the analysis reveals a negative correlation between social aspects and technological dominance. A statistically significant negative relationship exists between the desire for a positive learning environment and the belief that technology can replace human resources (τ = −0.181, p = 0.007). A similar negative correlation was found regarding the importance of feedback (τ = −0.171, p = 0.012). The more highly an individual values the emotional climate and human feedback, the more categorically they reject the possibility of technology replacing the educator. It is precisely the “learning atmosphere” that respondents consider the most difficult element to simulate using AI.

Table 8 summarizes the results of Kendall’s correlation analysis, examining the specific influence of demographic indicators—age, education, and computer skills—on the relationships under study.

The correlation analysis reveals that, in most cases, the respondents’ sociodemographic profile (age, education, computer skills) does not significantly influence their views on the importance of communication. However, one specific area was identified—learner autonomy—where statistically significant differences are observed. A statistically significant negative correlation was found between respondents’ age and the statement “I can effectively learn even without interaction with the teacher and other students” (τ = −0.161, p < 0.01). These results indicate a trend: the younger the respondent, the higher they rate their ability to learn autonomously. Conversely, as age increases, respondents become more skeptical about the possibility of qualitatively mastering the material without social interaction. This could be explained by the younger generation being more accustomed to independent information acquisition in the digital environment, while older learners link the learning process more closely to the presence of an educator.

A negative correlation was also fixed between level of education and autonomy (τ = −0.130, p < 0.05). Respondents with higher education are less convinced about learning without social contact. This likely occurs because higher education emphasizes the critical role of discussion and social networking in knowledge construction. Individuals with higher qualifications realize that “effective learning” is more than mere information consumption; therefore, they are less likely to agree that full learning can occur in isolation.

Crucially, computer skills showed no statistically significant correlation (p > 0.05) with any of the five criteria. This implies that an individual’s technological competence does not change their fundamental need for feedback, collaboration, and a positive environment. Even highly skilled IT users do not believe that technology could fully replace human resources (τ = 0.000). Similarly, there are no significant differences between demographic groups regarding the importance of feedback and a positive environment—these values remain universally important to everyone, regardless of age or education.

5. Discussion

The study results provide an in-depth understanding of the dynamics of AI integration in lifelong learning, emphasizing that technological adaptation is a complex process influenced less by demographic indicators and more by an individual’s skill and value system.

5.1. Resilience of Core Values Under Technological Transformation (RQ1)

In response to RQ1, the analysis of individual values according to S. Schwartz’s model revealed a high level of basic value stability. The study data do not support the assumption of an immediate transformative impact of AI on an individual’s moral and ethical foundation, as no statistically significant changes were observed in any of the ten studied value dimensions. Such resilience of the value system aligns with the theoretical insight by Dzalbe et al. (2015), stating that value stability serves as a societal adaptation instrument, providing the necessary psychological security during times of technological upheaval.

This stability aligns with the authors’ earlier findings on external crises (Balceraite & Lubkina, 2023; Balceraite et al., 2024). Unlike acute physical or economic threats that can temporarily prioritize ‘Security’ and ‘Tradition’, AI represents a technological transformation that does not trigger a significant shift in personal priorities. Consequently, core values consistently serve as a psychological safety anchor, preserving an individual’s internal integrity across entirely different types of societal transformations.

5.2. Digital Competence as the New Determinant of AI Perception (RQ2)

Addressing RQ2, one of the most significant findings of the study is the conclusion that respondents’ age is not a decisive factor in their overall attitude toward AI. Instead, the study data clearly identify digital skills as the primary factor determining the perception of AI’s utility and personal adaptation. This insight aligns with the findings of researchers (Prayuda et al., 2026), stating that digitalization today is primarily a matter of competence and skills.

Specifically, the digital divide no longer draws a boundary between age groups, but rather between functionally proficient and non-proficient users. While general acceptance is similar across generations, older respondents exhibit a more neutral and cautious stance, whereas the highest level of optimism is concentrated in the younger age groups. This suggests that policy interventions in lifelong learning should focus on skill acquisition rather than age-based segments.

5.3. The Socio-Centric Paradigm and Irreplaceability of Pedagogical Interaction (RQ3)

Regarding RQ3, the study highlights a clear “socio-centric” paradigm in lifelong learning, which is further supported by the research of other authors (Hachem et al., 2026; Prayuda et al., 2026). Although respondents acknowledge the potential of AI in personalizing materials, they express well-founded skepticism regarding AI’s ability to foster learner motivation. Respondents categorically reject the possibility that technology could fully replace human resources, confirming that learning is fundamentally a social process where the human element remains irreplaceable.

This polarization of views reinforces the role of AI in lifelong learning as a didactic aid, rather than a replacement for the educator. The findings emphasize that emotional support and quality feedback—elements that algorithms cannot fully simulate—remain the cornerstone of the pedagogical process. Academic education further fosters this critical view, as respondents with higher degrees express greater concern over the moral and ethical challenges posed by AI.

6. Limitations and Future Research

One of the most significant limitations of this study is related to the demographic structure of the respondent sample, which affects the representativeness of the data and the generalizability of the results. A further limitation of the study is its retrospective nature; the data reflect subjective perception dynamics rather than a real time series assessment. In future research, it would be advisable to conduct a classic longitudinal analysis over several years.

A pronounced gender imbalance is observed in the sample, where 89.6% (n = 144) of the respondents are female and 10.4% (n = 16) are male. Although this distribution precludes a full comparative analysis between genders and poses a risk of reflecting a specifically ‘feminine’ perspective on social interaction, it aligns with national statistical trends in Latvia, where females constitute the vast majority of graduates at all levels of higher education (Centrālā Statistikas Pārvalde, 2024). Consequently, the results reflect a highly educated demographic with advanced self-reflection and digital literacy. This specific profile likely accounts for the high degree of core value stability observed, as higher education serves as a significant factor in maintaining cognitive resilience during technological transitions.

Second, the sample is characterized by a high level of academic education (with over 70% holding a university degree), which may influence the results. Individuals with higher educational backgrounds often exhibit higher self-esteem and a more positive outlook on lifelong learning and technology adaptation. Consequently, the observed value stability and positive attitude toward AI integration might be more characteristic of this specific demographic group rather than the general population.

Identifying that the sample is dominated by respondents with higher education, it can be concluded that the obtained data accurately characterize the views of the most educated segment of society regarding AI. This limitation can also be interpreted as an advantage, as it allows the study results to be analyzed as a reasoned forecast of AI’s future impact on education, particularly from the perspective of individuals with experience in the academic environment. Thus, the study provides significant insight into how the most educated portion of society perceives this technology’s potential.

The data collection method and the principle of voluntary participation likely attracted respondents who already possess an interest in technology or lifelong learning, thereby excluding the less digitally active part of society.

Despite these limitations, the study provides valuable insight into the attitudes of the most educated and active segment of society in lifelong learning (primarily consisting of women), which represents a crucial target audience for educational policy-making. Future research should utilize a stratified sample, intentionally increasing the proportion of male respondents and those with secondary education to achieve a more balanced overall perspective.

Furthermore, this study identifies potential directions for future research. The observation that some respondents declined to participate due to a perceived lack of knowledge highlights the need for an in-depth investigation into public understanding of AI. This provides a rationale for new studies employing a methodology that includes an initial educational phase, ensuring a broader and more representative respondent sample.

7. Conclusions

Based on the empirical results of the study and the statistical data analysis, the following conclusions are drawn.

Regarding RQ1, empirical evidence confirms that an individual’s core value system remains highly stable despite the rapid integration of AI. No significant shifts were observed in any of the ten Schwartz value dimensions. This resilience suggests that core values act as a crucial psychological anchor during technological stress, preserving internal integrity while individuals acquire new digital skills.

In response to RQ2, the data demonstrates that the decisive role in the perception of artificial intelligence is played by the individual’s level of digital skills rather than age, thus debunking stereotypes regarding an inevitable “generation gap” in technology adaptation processes. Respondents with higher digital competence show a more direct link between the opportunities provided by technology and personal adaptation, while academic education forms a critical attitude towards the ethical risks posed by artificial intelligence. At the same time, the attitude toward the role of AI specifically in the learning process is universal across all demographic groups, where the potential for content individualization is highly valued. However, serious concerns remain regarding technological dependency, which received the highest average rating in the entire survey.

For RQ3, the study highlights a dominant socio-centric paradigm in lifelong learning. Although AI is recognized as a powerful didactic tool, participants overwhelmingly agree that algorithms cannot replace human resources. Direct communication, emotional support, and quality feedback remain irreplaceable, reaffirming that the human element is critical for sustaining motivation and fostering a positive educational environment in the AI era.