UTILIZATION OF ARTIFICIAL INTELLIGENCE TOOLS AMONG MID-LIFERS IN THE GOVERNMENT SERVICE

Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [406] UTILIZATION OF ARTIFICIAL INTELLIGENCE TOOLS AMONG MID-LIFERS IN THE GOVERNMENT SERVICE Joseph Isaiah S.Trinidad Kimberly D. Tesoro Abbel F. Unla Erovenos L. Emata Gaudencio G. Abellanosa University of Southeastern Philippines, College of Development Management Graduate Program, Mintal Campus, Davao City ABSTRACT This study aimed to determine the significant difference in the utilization of Artificial Intelligence tools among mid-lifers in the government service. The researchers utilized non-experimental quantitative and comparative research design to determine significant differences. It is anchored on the Institutions specifically supporting the objective of Target 16.6 under this goal, develop effective, accountable and transparent institutions at all levels. The study was conducted with 105 employees from the different government agencies. The researchers found out that mid-lifers revealed moderate level as to the utilization of AI tools in terms of construct Perils. Moreover, they manifested high level for the constructs Promises and Powers. Lastly, they disclosed nonsignificant difference in the level of utilization of AI tools when they are grouped by gender, work, status, and length of service in terms of constructs Perils and Promises. Mid-lifers manifested significant difference in the utilization of AI tools, particularly on the indicator Perils on which mid-lifers handling Permanent position is leading. No significant difference was noted for the indicators Promises and Power. Keywords: Artificial Intelligence, Mid-Lifers, Government Service, Perils, Promises, Power INTRODUCTION In the context of public administration, employees’ ability to understand and utilize AI tools is critical to successful digital transformation. Mid-lifers, typically aged 35–55, occupy essential operational and supervisory roles that directly influence the adoption and implementation of emerging technologies in government service (AP-NORC, 2024). Unlike younger digital natives, mid-lifers often balance substantial institutional experience with the need to adapt to rapidly evolving technological landscapes. Research shows that this age group may exhibit distinct perceptions toward AI—ranging from enthusiasm and perceived usefulness to concerns about ethical risks, job displacement, and system reliability (Wong et al., 2025). Contemporary studies highlight that employee attitudes toward AI adoption can be understood through three major dimensions: Perils (perceived risks), : Promises (perceived benefits), and Power (perceived capability of AI systems), as conceptualized by Shum and Lau (2024). High levels of perceived promises are often associated with improved performance, reduced workloads, and enhanced service quality (Huang & Rust, 2021). Conversely, perceived perils stem from fear of surveillance, data breaches, ethical issues, and job insecurity (Gerlich et al., 2023; UNESCO, 2021). Perceived power relates to AI’s increasing ability to equal or exceed human performance in routine or analytical tasks (Zhang & Lu, 2023). Within the Philippine government setting, interest in AI integration has grown, particularly in areas related to administrative processing, public employment services, frontline transactions, and data-driven decision-making. However, empirical studies that directly assess AI utilization among mid-life government employees remain scarce. Understanding their perceptions and utilization patterns is essential because mid-lifers often serve as digital transition leaders and mentors within their organizations. Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [407] Given this context, the present study investigates the utilization of AI tools among mid-lifers in government service. It examines their perceived promises, perils, and power of AI and analyzes whether significant differences exist across demographic characteristics such as gender, work type, employment status, and length of service. This inquiry offers crucial insights into the readiness of the mid-life public workforce to adopt AI technologies, thereby supporting the broader goal of fostering effective and technologically adaptive government institutions. OBJECTIVES This study aimed to determine the significant difference in the utilization of Artificial Intelligence tools among mid-lifers in the government service. Specifically, it aims to: 1) Determine the demographic profile of the respondents. 2) Level of utilization of AI tools among mid-lifers; 3) Significant difference the level of utilization of AI tools among mid-lifers when they are grouped by gender, work, status and length of Service. METHODOLOGY This research constitutes a comparative analysis, a methodical examination that directly compares two or more things to identify their similarities and differences (Jean Kaluza, 2023). The primary aim of the study is to ascertain significant differences in the level of utilization of AI tools among mid-lifers based on their demographic profiles. Data for this study were gathered through an adaptive questionnaire from “Perils, power and promises: Latent profile analysis on the attitudes towards artificial intelligence (AI) among middle-aged and older adults in Hong Kong” (Shum et al., 2024). The study involved 105 respondents from the different government agencies. The questionnaire comprised two distinct sections. The first section captured demographic characteristics, including gender, work, status and length of service. The second section consisted of 20 items across three indicators. The following statistical tools were used such as mean, t-test, and ANOVA. RESULTS AND DISCUSSION This chapter presents the results of the study, including the descriptive statistics (mean and standard deviation), along with the corresponding interpretations and analyses. Tables are utilized to illustrate the findings, and the discussion of both tabular and graphical data is provided to facilitate clarity and understanding. Demographic Profile Presented in Table 1 is the demographic profile of the respondents. There were 105 respondents, 52 male and 53 female equivalent 49.5% and 50.5% respectively. Most of the respondent are male. There were 5 respondents occupying teaching position equivalent to 4.8 percent while non-teaching was 100 equivalent to 95.2 percent. Most of the respondents were non-teaching positions. Furthermore, out of 105, 71 or 67.6 percent of the respondents hold a permanent status while 32.4 percent or 34 respondents equivalent to 74.3 percent are nonpermanent. Lastly, regarding the length of service, 78 respondents served for less than 10 years, 19 respondents served for 11-20 years while 8 respondents served for over 21 years. Most of the respondents served for less than 10 years. This finding reflects broader trends observed in public-sector employment in recent years. The nearly equal distribution of male and female respondents aligns with global patterns showing that government institutions are among the most gender-balanced workplaces (OECD, 2023). The predominance of non-teaching and administrative personnel in the sample is consistent with national and international reports indicating that most mid-level government functions are occupied by technical and administrative staff rather than instructional personnel (AP-NORC, 2024). The high proportion of permanent employees likewise mirrors public-sector workforce structures in Southeast Asia, where permanent appointments remain the most common employment status in national government agencies (Asian Development Bank, 2021). Additionally, the large number of respondents with fewer than ten years in service supports findings that many governments have expanded hiring efforts in response to modernization initiatives, digital transformation demands, and evolving public service needs (World Bank, 2024). These demographic patterns correspond with studies noting that mid-lifers often possess diverse lengths of service as government agencies undergo transitions toward more technology-enabled systems (Duck & Ernest 2025). Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [408] Indicator Frequency Percent Gender Male 52 49.5 Female 53 50.5 Total 105 100 Type of Work Teaching 5 4.8 Non-Teaching 100 95.2 Total 105 100 Status Permanent 71 67.6 Non-Permanent 34 32.4 Total 105 100 Length of Service 1-10 78 74.3 11-20 19 18.1 21 above 8 7.6 Total 105 100 Table 1. Demographic Profile Utilization of AI Tools Among Mid-lifers Perils. Presented in Table 2 is the level of AI tools among mid-lifers. For the indicator perils, the respondents revealed moderate level of the means score of 3.40. The mid-lifers claimed that artificial intelligence is used to spy on people with the mean score 3.66 or high level. The mid-lifers claimed that they sometimes shiver with discomfort when they think about future uses of Artificial Intelligence with mean score of 3.19 or moderate level. Further, data revealed that the mid-lifers claimed that artificial intelligence often takes control of people, threatening, use to spy people and take control of people. They sometimes claim that AI poses danger when thinking about its future uses. Moreover, mid-lifers sometimes think that they will suffer if AI is utilized unethically and that it may sometimes commit frequent errors. Promises. As shown in the table below, mid-lifers often perceive that AI is exciting with a mean score of 3.97 or high level of utilization. While they perceive that for routine transactions, they would rather interact with an artificially intelligent system than with a human. The mid-lifers find the idea of AI exciting because it can be used in their current position as it often provides new economic opportunities and provides positive impacts so that the future of society will benefit from AI utilization. Lastly, for routine transactions, mid-lifers would rather interact with an AI system than with a human. Power. As exhibited in the table, the respondents showed a high level of appreciation for the beneficial applications of AI with a high level of utilization, and mean score of 4.04. Additionally, mid-lifers often think that AI has the potential to outperform humans in routine jobs and convey their overall impressiveness with what AI can do. Utilization of AI Tools SD Mean Descriptive Level Perils 1.02 3.4 Moderate I think Artificial Intelligence is dangerous. 1.25 3.41 Moderate Artificial Intelligence might take control of people. 1.22 3.58 High I find Artificial Intelligence threatening. 1.24 3.5 High I shiver with discomfort when I think about future uses of Artificial Intelligence. 1.41 3.19 Moderate People like me will suffer if Artificial Intelligence is used more and more. 1.48 3.28 Moderate Organizations use Artificial Intelligence unethically. 1.17 3.34 Moderate Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [409] Artificial Intelligence is used to spy on people. 1.28 3.66 High I think artificially intelligent systems make many errors. 1.11 3.24 Moderate Promises 0.78 3.76 High Artificially intelligent systems can help people feel happier. 1 3.74 High I am interested in using artificially intelligent systems in my daily life. 1.12 3.72 High Artificial Intelligence is exciting. 0.98 3.97 High Much of society will benefit from a future full of Artificial Intelligence. 0.92 3.86 High I would like to use Artificial Intelligence in my own job. 0.96 3.84 High Artificial Intelligence can provide new economic opportunities. 0.95 3.95 High Artificial Intelligence can have positive impacts on people's wellbeing. 0.98 3.7 High For routine transactions, I would rather interact with an artificially intelligent system than with a human. 1.21 3.28 Moderate Power 0.77 3.72 High There are many beneficial applications of Artificial Intelligence. 0.89 4.04 High An artificially intelligent agent would be better than an employee in many routine jobs. 1.06 3.54 High Artificial Intelligent systems can perform better than humans. 1.22 3.56 High I am impressed by what Artificial Intelligence can do. 1.1 3.72 High Table 2. Level of Utilization of AI Tools Among Mid-lifers The patterns observed in the respondents’ perceptions of AI, particularly the moderate concerns related to perils and the strong acknowledgment of promises and power are consistent with recent empirical research. Several studies show that individuals, especially mid-career and older workers, frequently express caution about AI due to perceived risks such as surveillance, data misuse, algorithmic errors, and loss of control (Gerlich et al., 2023; UNESCO, 2021). These concerns mirror the respondents’ moderate agreement that AI can be threatening, may commit errors, and could lead to unethical applications if not properly regulated. At the same time, literature from 2020 to 2025 highlights that workers also recognize the substantial benefits of AI, including efficiency gains, improved public service delivery, and enhanced job performance (Huang & Rust, 2021; Wong et al., 2025). This supports the high mean scores in the Promises construct, where mid-lifers view AI as exciting, helpful for routine tasks, and capable of creating economic and organizational value. Furthermore, the strong agreement under the Power construct aligns with global findings demonstrating that employees increasingly believe AI can outperform humans in routine and repetitive functions due to its analytical and computational strengths (Zhang & Lu, 2023). Overall, these studies validate the respondents’ mixed but generally positive attitudes toward AI, reflecting both cautious optimism and recognition of AI’s expanding role in government work. Utilization of AI Tools when Grouped by Gender Perils. Presented in Table 3 is the non-significant difference in the level of utilization of AI tools among midlifers when they are grouped in gender. The mid-lifers manifested non-significant difference as the indicator perils reflected in the t-value of 0.79 with the p-value of 0.43 which is less than 0.05 level of significance. The result is not significant and the acceptance of the null hypothesis. This implies that male and female manifested equal perception as to the utilization of AI tools along this indicator. Promises. As shown in the table, mid-lifers disclosed non-significant differences on the construct promise which then showed a t-value of 1.64 with the p-value of 0.10 which is less than 0.05 level of significance. The result is not significant hence the null hypothesis is accepted. This infers that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of promises. Power. Indicated in the table, mid-lifers conveyed non-significant difference on the construct power which showed a t-value of 0.69 with the p-value of 0.49 which is lesser than 0.05 level of significance. The result is not significant thus the null hypothesis is accepted. This indicates that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of power. Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [410] Utilization of AI Tools Male Female t-value p-value Decision on H0 Perils 3.48 3.32 0.79 0.43 Accept Promises 3.88 3.63 1.64 0.10 Accept Power 3.77 3.67 0.69 0.49 Accept Table 3. Significant Difference in the Level of Utilization of AI Tools when Grouped by Gender No significant differences were found across all constructs, suggesting similar AI perceptions among male and female mid-lifers consistent with global findings (Zou & Schiebinger, 2022). Utilization of AI Tools when Grouped by Work Perils. Presented in Table 4 is the non-significant difference in the level of utilization of AI tools among midlifers when they are grouped by work. The mid-lifers manifested non-significant difference as the indicator perils reflected in the t-value of 0.34 with the p-value of 0.70 which is less than 0.05 level of significance. The result is not significant and the acceptance of the null hypothesis. This implies that male and female manifested equal perception as to the utilization of AI tools along this indicator. Promises. As shown in the table, mid-lifers disclosed non-significant differences on the construct promise which showed a t-value of 1.61 with the p-value of 0.11 which is lesser than 0.05 level of significance. The result is not significant hence the null hypothesis is accepted. This implies that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of promises. Power. Lastly, mid-lifers conveyed non-significant differences on the construct power which then showed a tvalue of -1.59 with the p-value of 0.12 which is lesser than 0.05 level of significance. The result is not significant; therefore, the null hypothesis is accepted. This implies that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of power. Utilization of AI Tools Teaching Non-Teaching t-value p-value Decision on H0 Perils 3.58 3.39 0.34 0.70 Accept Promises 4.30 3.73 1.61 0.11 Accept Power 4.25 3.69 1.59 0.12 Accept Table 4. Significant Difference in the Level of Utilization of AI Tools when Grouped by Work Utilization of AI Tools when Grouped by Status Perils. Disclosed in Table 5 is the significant difference in the level of utilization of AI tools among mid-lifers when they are grouped by status as shown in the t-value of 3.07 with a p-value of .000 which is lesser than 0.05 level of significance. The result is significant and the rejection of the null hypothesis. This implies that midlifers having permanent status showed higher level of utilization of AI tools compared to non-permanent status. The findings support the study of (Duck & Ernest 2025) as AI tools usage enhances employees’ work performance in key areas and it reduced burnout and lowered turnover rates. Promises. As shown in the table, mid-lifers disclosed non-significant differences on the construct promise which showed a t-value of 1.14 with the p-value of 0.26 which is less than 0.05 level of significance. The result is not significant hence the null hypothesis is accepted. This implies that mid-lifers having permanent and nonpermanent status, displayed equal level of perception towards the utilization of AI along with this indicator. Power. Mid-lifers conveyed non-significant difference in the construct power which then showed a t-value of 1.94 with the p-value of 0.06 which is lesser than 0.05 level of significance. The result is not significant therefore the null hypothesis is accepted. This implies that mid-lifers occupying permanent and non-permanent positions showed equal level of perception towards the utilization of AI for this indicator. Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [411] Utilization of AI Tools Permanent Non-Permanent t-value p-value Decision on H0 Perils 3.61 2.98 3.07 0.00 Reject Promises 3.82 0.82 1.14 0.26 Accept Power 3.82 0.83 1.94 0.06 Accept Table 5. Significant Difference in the Level of Utilization of AI Tools when Grouped by Status Utilization of AI Tools when Grouped by Length Perils. Presented in Table 6 is the non-significant difference in the level of utilization of AI tools among midlifers as classified by work. The mid-lifers manifested non-significant difference as the indicator perils reflected in the f-value of 1.85 with the p-value of 0.162 which is lesser than 0.05 level of significance. The result is not significant and the acceptance of the null hypothesis. This implies that male and female manifested equal perception as to the utilization of AI tools along this indicator of perils. Promises. Furthermore, mid-lifers disclosed non-significant differences on the construct promise which then showed a f-value of 0.524 with the p-value of 0.594 which is lesser than 0.05 level of significance. The result is not significant hence the null hypothesis is accepted. This implies that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of promises. Power. Finally, mid-lifers conveyed non-significant difference on the construct power which then showed a fvalue of 1.34 with the p-value of 0.27 which is lesser than 0.05 level of significance. The result is not significant hence the null hypothesis is accepted. This implies that male and female mid-lifers, showed equal levels of perception towards the utilization of AI for the indicator of power. Utilization of AI Tools 1-10 11-20 21-above F-value p-value Decision on H0 Perils 3.30 3.80 3.46 1.85 0.162 Accept Promises 3.72 3.83 3.99 0.524 0.594 Accept Power 3.64 3.92 3.94 1.34 0.267 Accept Table 6. Significant Difference in the Level of Utilization of AI Tools when Grouped by Length of Service CONCLUSION Overall, mid-lifers in government service demonstrate a generally positive perception of AI tools, especially in terms of their usefulness, efficiency, and capacity to enhance organizational performance. This aligns with global findings that public-sector employees recognize AI’s potential to streamline administrative processes and improve decision-making (World Bank, 2024; Wong et al., 2025). Although concerns about risks such as privacy issues, algorithmic errors, and ethical misuse remain moderate, these apprehensions are common among mid-career workers who are adapting to rapid technological changes (Gerlich et al., 2023). The minimal differences across demographic categories suggest that attitudes toward AI among mid-lifers are relatively consistent, indicating a generally receptive workforce capable of supporting digital governance initiatives. However, the significant variation observed among permanent employees implies that those with long-term roles may feel more accountable for potential consequences of AI misuse and thus express heightened caution. Studies show that permanent staff often perceive greater institutional responsibility and therefore require more structured guidance in navigating emerging technologies (Duck & Ernest 2025). These findings underscore the need for continuous AI-related capacity-building programs that emphasize not only technical skills but also responsible, ethical, and transparent use of AI tools, as recommended by (UNESCO, 2021) and (OECD, 2023). By addressing both confidence and concerns, government agencies can better prepare their mid-life workforce to engage with AI innovations effectively and responsibly. RECOMMENDATION Based on the findings of the study and supported by recent research on public-sector AI adoption, the following recommendations are proposed: Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [412] 1. Strengthen AI Literacy and Skills Training for Government Mid-Lifers. Given that mid-lifers demonstrated strong interest in the benefits and capabilities of AI, government agencies should institutionalize regular capacity-building programs that focus on practical AI usage, data literacy, and integration into administrative workflows. Studies show that AI adoption is significantly higher when employees receive structured training and organizational support (Wong et al., 2025; World Bank, 2024). Training should include hands-on sessions involving tools already used in public offices, such as automated text generators, data analytics platforms, and AI-driven public service applications. 2. Implement Ethical and Responsible AI Use Workshops. Because moderate concerns about AI-related risks were observed, it is recommended that agencies incorporate ethics-focused training that addresses algorithmic bias, data privacy, transparency, and responsible use of automated systems. UNESCO (2021) emphasizes that public-sector workers must be trained to understand the ethical implications of AI, while OECD (2023) stresses the importance of ethical governance frameworks. Workshops should include real-world case examples to help employees make informed decisions when using AI tools. 3. Provide Targeted Support for Permanent Employees Showing Higher Perceived Risks. Permanent employees exhibited significantly higher concerns under the “Perils” construct. This indicates a need for targeted interventions that address their elevated sense of responsibility and accountability. According to (Duck & Ernest 2025), employees with longer tenure or permanent roles tend to be more cautious toward digital innovations due to concerns about long-term job implications and institutional risks. Tailored mentoring, coaching, and open dialogues on AI governance can help reduce apprehensions. 4. Institutionalize Clear AI Governance Policies Across Government Agencies. To address concerns about surveillance, misuse, and algorithmic errors, agencies should develop or adopt standardized AI governance guidelines. These may include protocols on data privacy, transparency, accountability, and human oversight. Global frameworks such as UNESCO's Recommendation on the Ethics of Artificial Intelligence (2021) and OECD’s AI Principles (2023) can serve as templates for local policy formulation. 5. Promote a Culture of Human - AI Collaboration. Since mid-lifers already perceive AI as useful and capable, agencies should encourage work models that highlight AI as a supportive tool rather than a replacement for human workers. Research shows that workers adapt more positively when AI is framed as a collaborative system that enhances human decision-making rather than replacing roles (Gerlich et al., 2023; Huang & Rust, 2021). Leadership communication should reinforce this perspective to build trust and confidence in AI technologies. 6. Conduct Continuous Monitoring and Evaluation of AI Utilization in Government Service. To ensure effective adoption, agencies should regularly assess how AI tools are used, identify challenges, and gather feedback from employees. Continuous evaluation helps align AI initiatives with organizational goals and workforce needs, as recommended by the World Bank (2024) and AP-NORC (2024). This monitoring can guide future policy adjustments and training programs. ACKNOWLEDGEMENT We would like to extend our sincere gratitude to Dr. Gaudencio G. Abellanosa, for his invaluable assistance and guidance throughout our research endeavor. His expertise and support have been instrumental in shaping the direction and outcomes of our study. We deeply appreciate his dedication and commitment to helping us navigate through the intricacies of our research process. REFERENCES [1] AP-NORC Center for Public Affairs Research. (2024). Workplace technology and attitudes of mid-career professionals. AP-NORC. Volume-09 Issue 11, November-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [413] [2] Asian Development Bank. (2021). Public sector management report: Workforce structures in Southeast Asia. ADB Publications. [3] Duck, A. S., & Earnest, D. (2025). Impact of Artificial Intelligence on Employee Job Performance, Communication, and Job Security. [4] Gerlich, R., Baumann, A., & Hunning, T. (2023). AI anxiety in the workplace: Antecedents and consequences. Journal of Business Research, 161, 113897. https://doi.org/10.1016/j.jbusres.2022.113897 [5] Huang, M.-H., & Rust, R. (2021). A strategic framework for artificial intelligence in marketing. Journal of the Academy of Marketing Science, 49, 30–50. https://doi.org/10.1007/s11747-020-00730-0 [6] OECD. (2023). Public employment outlook 2023. OECD Publishing. [7] Shum, A., & Lau, J. (2024). Perils, power, and promises: Latent profile analysis on attitudes toward artificial intelligence among middle-aged and older adults. Computers in Human Behavior, 152, 107154. https://doi.org/10.1016/j.chb.2023.107154 [8] UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO Publishing. https://unesdoc.unesco.org/ [9] Wong, M., Chiu, S., & Ng, E. (2025). Determinants of AI adoption intention among government workers in Asia. Asia Pacific Journal of Public Administration, 47(1), 1–19. [10] World Bank. (2024). Government AI readiness report 2024. World Bank Publications. https://www.worldbank.org/ [11] Zhang, T., & Lu, Y. (2023). Perceived benefits and risks of AI-enabled services: A systematic review. Information Systems Frontiers, 25, 765–780. https://doi.org/10.1007/s10796-022-10343-9 [12] Zou, J., & Schiebinger, L. (2022). Ensuring gender-fair AI for public service workplaces. Nature Machine Intelligence, 4, 986–995.