Developing and Validating the AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0): Initial Evidence from Architecture Design Education in Pakistan

Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 303 ©2025 PJSS, Bahauddin Zakariya University Multan Pakistan Developing and Validating the AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0): Initial Evidence from Architecture Design Education in Pakistan a Muhammad Aarez Ali, b Malik Javied Anwar a Assistant Professor, Beaconhouse National University, Lahore, Pakistan Email: a[email protected].pk b Assistant Professor, Superior University, Lahore, Pakistan Email: javied.anwa[email protected] ARTICLE DETAILS ABSTRACT History: Accepted: 05 December, 2025 Available Online: 17 December, 2025 Purpose: This study develops and provides pilot validation for the AI Readiness and 21stCentury Skills Scale for Industry 5.0 (AIRS-5.0), a multi-dimensional instrument designed to measure how traditional architecture curriculum and AI literacy equip graduates with Industry 5.0 readiness through AI-enabled 21st-century skills. Design/Methodology/Approach: The scale integrates four core constructs: 21st-century skills acquired through traditional curriculum (TC_21C), AI Literacy (AILS), AI-Enabled 21stCentury Skills (AI_21C), and Industry 5.0 Readiness (IR5). Following expert review and iterative refinement, the 91-item instrument was pilot-tested among 33 final-year architecture students and 14 thesis tutors from ten PCATP-accredited institutions. SPSS v26 and PROCESS Macro Model 4 were used to test reliability and preliminary mediation analyses. Given the small sample size, the findings are exploratory and not intended for generalization. Findings: Results obtained were extremely promising with excellent internal reliability (Cronbach α = 0.95 – 0.99). However, the high MIIC values (above 0.60) outlined some risk of redundancy. AI Literacy (AILS) emerged as the strongest predictor of Industry 5.0 readiness (β = 0.67, p < 0.001), while the mediational effect of AI_21C, a novel construct, was insignificant. Tutors’ assessment of students’ readiness across major AI-related constructs were significantly lower than students’ self-reported values, highlighting a gap in AI engagement perception. Overall, the findings validate the scale’s structure as well as its potential to articulate pedagogical gaps in AI integration across architecture curriculum. Implications/Originality/Value: The AIRS-5.0 scale provides a holistic foundation for mapping AI readiness and 21C competencies in design education. By linking architectural pedagogy, AI literacy, and Industry 5.0’s human-centric ethos, the study offers a rare empirical framework for guiding curriculum reform, faculty development, and institutional AI integration in creative disciplines. © 2025 The authors. Published by PJSS, BZU. This is an open-access research paper under the Creative Commons Attribution-Non-Commercial 4.0 Keywords: Industry 5.0 21st-Century Skills AI Literacy Architecture Education Scale Development Recommended Citation: Ali, M. A., & Anwar, M. J. (2025). Developing and Validating the AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0): Initial Evidence from Architecture Design Education in Pakistan. Pakistan Journal of Social Sciences, 45(4), 303-318. DOI: 10.5281/zenodo.17959675 *Corresponding Author’s email address: [email protected]u.pk Pakistan Journal of Social Sciences ISSN (E) 2708-4175 ISSN (P) 2074-2061 Volume 45: Issue 4 December 2025 Journal homepage: https://pjss.bzu.edu.pk Pakistan Journal of Social Sciences ISSN (E) 2708-4175 ISSN (P) 2074-2061 Volume 45: Issue 4 December 2025 Journal homepage: https://pjss.bzu.edu.pk Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 304 1. Introduction The shift between Industry 4.0 to Industry 5.0 is altering the way human beings and technology interact. Industry 4.0 emphasized automation and the efficiency of data, but Industry 5.0 returned the human factor and appreciated creativity, ethical standards, teamwork, and sustainability through the partnership of people and intelligent systems (Gobakhaloo, Mahdiraji, Iranmanesh, & Jafari-Sadeghi, 2024; European Commission, 2021). Preparation for this stage requires graduates that combine technical and digital fluency with critical thinking, communication skills, and adaptability. These skills are what are perceived as the 21st-century skills in major global frameworks (Binkley, et al., 2010; Battelle for Kids, 2019; UNESCO, 2023; OECD, 2018). 21st-century skills are now defined as a set of skills that are necessary for success in today’s complex and rapidlyevolving digital-centric world. Voogt and Roblin (2012) divide these skills into three main domains: Cognitive skills include capabilities like creativity and critical judgment; interpersonal skills allude to strong communication and leadership skills, while key intrapersonal skills include adaptability, ethical responsibility, and readiness for lifelong learning (Binkley, et al., 2010). However, these generic frameworks typically overlook discipline-specific 21C-skills. It is thus understood that these skills are not limited to any one profession; they are a key requirement for any graduate entering a rapidly digitizing industry. Educational institutions worldwide are now looking to update their pedagogies to embed these ideals into their curriculum to prepare their graduates for Industry 5.0 (Voogt & Roblin, 2012; OECD, 2018; European Commission, 2021). Education in creative practices, and especially architecture already cultivates many of these skills, though these skills remain tacit and rarely measured or assessed. The iterative and reflective nature of studio-based architectural education already embeds most of these 21stcentury skills like creativity, adaptability, communication, and ethical judgement through learning-by-doing (Schon, 1983; Salama, 2015). The cyclic nature of making, criticizing, reflecting, and re-doing is an adequate reflection of solving real-world problems and collaboration (Ceylan, Sahin, Secmen, Somer, & Suher, 2020). However, the primacy of technical and design skills remains; these competencies are understood to be nurtured automatically. They are never measured or assessed independently (Khodeir & Nessim, 2020). Recent global trends such as rise of Building Information Modeling (BIM) and Artificial Intelligence (AI) have pushed architecture schools towards a more digital, collaborative, and AI-enabled workflows and environments (Asghar, 2023; Basarir, 2022). In Pakistan, both the Higher Education Commission (HEC) and Pakistan Council of Architects and Town Planners (PCATP) have recognized the need and pushed for outcome-based, digitally integrative curriculum (NCRC - Architecture, 2025). This changing environment necessitates multi-domain survey instruments like AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0) that capture the relationship between traditional curriculum, 21C skills, AI literacy, and Industry 5.0 readiness. The last few years have seen an influx of AI tools within educational and creative landscape. Wicks and Paulus (2022) argue that these intelligent, data-based analytical tools have expanded the scope of the originally envisioned 21C skills. This research builds upon these ideas to develop a case for how these cognitive, interpersonal, and intrapersonal skills are being reshaped into human-AI hybrid capabilities; herein referred to as AI-enabled 21stcentury skills (AI_21C). Industry 5.0’s vision for a human-centric, technology driven future naturally leads to such a concept (European Commission, 2021; An, 2024; Woo, 2025). However, existing studies tend to address these ideas in isolation. Some focus entirely on 21C skills (Khodeir & Nessim, 2020), whether others frame AI literacy (Wang, Rau, & Yuan, 2022), without linking them to Industry 5.0 readiness. Similarly, other studies, both local (Abro & Nazir, 2024) and international (Maina & Daful, 2017; Salleh, Md Yusof, & Memon, 2016) have studied the gap between academia and practice without linking those to AI literacy and development of 21st-century skills. There is no validated scale that captures the intersection of traditional curriculum, 21st-century skills, AI literacy, and readiness for industry 5.0 in an industry that is rapidly evolving towards human-AI collaborative, sustainable professional environment. Therefore, this study proposes a cohesive measurement instrument, AI Readiness and Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 305 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0), to bridge this gap and provide a statistically validated tool to measure the graduates’ perceived readiness for Industry 5.0 through the lens of traditional curriculum, 21C skills, AI literacy, and AI-enabled 21C skills. This paper present Phase I of the development and validation of the scale, focusing on establishing content validity and internal consistency through expert review and pilot testing. 2. Theoretical Framework and Literature Review 2.1 21st-Century skills cultivated through traditional curriculum (TC_21C) The 21st-century skills frameworks, designed and refined to promote essential graduate competencies classify these skills into cognitive, interpersonal, and intrapersonal domains, encompassing creative, collaborative, and selfdevelopment capabilities. Originally designed for 21st-century education, this COG-INT-INTRA taxonomy (Binkley, et al., 2010) now aligns with Industry 5.0’s human-technology collaborative ethos (European Commission, 2021; Battelle for Kids, 2019; OECD, 2018). Architecture pedagogy, based primarily on its studio culture, automatically looks to cultivate many of these capabilities through the iterative process of making, critique, and reflecting, automatically teaching students the required capabilities for real-life problem-solving through decision-making, collaborating, and communicating their ideas (Schon, 1983; Salama, 2015). However, most architecture programs focus exclusively on technical and design quality in their assessment criteria; the aforementioned capabilities tacit and unassessed (Khodeir & Nessim, 2020; Ng, Mari, & Lin, 2022). In the proposed AIRS-5.0 framework, TC_21C is understood as the baseline 21C skills along the COG-INT-INTRA taxonomy developed through the core knowledge areas of the traditional, analog B. Arch curriculum. In Pakistan, these knowledge areas are defined by HEC (2025) and PCATP across design, technology, theory, and professional practice. These competencies are the foundation on which AI-enabled 21C skills are built upon. 2.2 A.I Literacy (AILS) AI literacy is the combination of knowledge, skills, and critical awareness required to work with intelligent systems (Modh, Bhatt, Kalyani, & Jain, 2025). International frameworks attribute four basic dimensions to this concept: Awareness, usage, evaluation, and ethics (Long & Magerko, 2020; Wang, Rau, & Yuan, 2022). These dimensions illustrate that digital literacy are expanded by critical evaluation and moral responsibility. Generative, analytical, and simulation-based AI tools are being increasingly used in architecture by both professionals and students. AI Literacy Scale (AILS) equips students to understand intelligent data-based systems and apply them creatively and critically within design process as a creative and analytical collaborator (Basarir, 2022; Suleman, 2024). Within the AIRS-5.0 framework, AILS is understood as a core independent construct, complementing TC_21C. Together, the traditional curriculum and the newly developed AI literacy help students develop AI-Enabled 21C skills (AI_21C), which in turn leads to enhanced Industry 5.0 Readiness (IR5) (Wicks & Paulus, 2022; European Commission, 2021). 2.3 Digitization of Global Architecture Curriculum – From Analog to AI-Enabled Learning Global higher education systems have undergone a swift digitization process. From incorporating automation and data driven tools, they have moved towards integrating intelligent systems and data-analytics into pedagogical streams. This shift is also advocated by international framework like European Commission (2021) and OECD (2020). Architecture education has also undergone this shift; first by integrating precision and efficiency based software like Computer-Aided Design (CAD), but Building Information Modelling (BIM) marked the first real Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 306 change by allowing structure, materials, construction technology, and performance data to coexist in one shared model (Bayhan & Karaca, 2020). The shift to online teaching mode during COVID-19 accelerated this shift. Public and private institutions alike had to switch towards online studios. Collaborative tools like Miro and Pildat demonstrated how online media can be used to continue a studio conversation and peer critique during the pandemic (Ceylan, Sahin, Secmen, Somer, & Suher, 2020; Bayhan & Karaca, 2020). This online shift made students and tutors alike more accustomed to digital learning and further paved the way for burgeoning AI tools to be understood as a tool for learning. However, research shows that the trend for AI adoption is similar to BIM; it is understood as a technical element used for efficiency rather than as a component of design thinking (Gabr, Elgendy, & Elmasry, 2022; Basarir, 2022). Students use AI tools primarily for visualization or efficiency but not to reason or reflect (Lu, Zhu, Pang, & Shadiev, 2024). However, educators argue that digital learning does not rely solely on technical ability, but also on judgment and morals. Introducing digital tools but not altering the method of teaching results in students being acquainted with software operation rather than reasoning (Milovanovic & Gero, 2020; Tedjosaputro, 2019). Recent research demands such a shift as a transition towards a model of literacy that combines comprehension, practice, and contemplation to fill this gap (Modh, Bhatt, Kalyani, & Jain, 2025; Basarir, 2022). Building upon these ideas, emerging frameworks position AI tools as a powerful force towards reshaping human capabilities (Long & Magerko, 2020; Wicks & Paulus, 2022). Scholars have introduced terms like Human-AI Collaboration (Woo, 2025), HumanAI skills (An, 2024), and AI-Enabled Human Competencies (BraiNet Journal Editorial Board, 2024). These developing frameworks indicate an evolution from tool-based AI usage towards hybrid human-AI capabilities. 2.4 AI-Enabled 21st-Century Skills (AI_21C) Introduced in this study, AI-Enabled 21st-Century Skills (AI_21C) is a novel idea that can be understood as the next step in the evolution of 21C competencies in the age of AI and intelligent systems. Inherent human qualities like creativity, critical thinking, adaptability, collaboration, and communication with others can be expanded and expressed through digital and intelligent tools (Wicks & Paulus, 2022; BraiNet Journal Editorial Board, 2024). Through combining human intuition and ethical reasoning with data-based analysis and optimization processes, these hybrid skills emerge through human-AI collaboration in generating, testing, and refining design ideas (Ng, Mari, & Lin, 2022). In architecture design, generative design and simulation workflows provide the most direct expression for these skills. Students enhance their analytical and design skills by moving from purely human-driven work process towards a more integrative approach using digital and AI tools. In doing so, they shift from using technology more as a collaborator than purely as a tool (Basarir, 2022; Suleman, 2024). In this study, AI-enabled 21C skills (AI_21C) acts as a mediator that translates TC_21C and AILS into Industry 5.0 Readiness (IR5). It brings into context the shift that architectural education requires to prepare its graduates for the hybrid human-AI ethos of Industry 5.0 (European Commission, 2021). In the model, AI_21C. These skills are also understood within the context of the core knowledge areas of architecture education. 2.5 Industry 5.0 Readiness (IR5) Industry 5.0 ethos propose an evolution from the automation and efficiency-centered foci of Industry 4.0 towards a more human-centric view of technology and innovation. It values the human values of creativity, critical reflection, ethical judgment, and collaboration with intelligent systems as key towards a more resilient, sustainable, and innovative future (European Commission, 2021). Readiness in this context refers to graduates’ preparation for this rapidly-evolving human-AI collaborative professional environment (Gobakhaloo, Mahdiraji, Iranmanesh, & Jafari-Sadeghi, 2024). Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 307 The core dimensions of Industry 5.0 Readiness (IR5), as derived from the ethos of Industry 5.0 and validated scales such as Work Readiness Scale (WRS) by Caballero, Walker, and Fuller-Tyszkiewicz (2011) are  Human-AI collaboration and Ethics: working productively with intelligent systems while maintaining human ethical judgment  Sustainability and Systems Thinking: integrating environmental, social, and holistic perspectives in decision-making  Teamwork and Professionalism: adapting to cross-disciplinary environment  Career Readiness and Lifelong Learning: engaging in continuous skill upgrading in a rapidly-changing digital context (European Commission, 2021) The evolving direction of architecture industry demands graduates who can merge human creativity and ethical mindset with AI-assisted workflows towards designing a more resilient, sustainable future (Suleman, 2024). Industry 5.0 Readiness (IR5) thus serves as the dependent variable in the AIRS-5.0 framework, assessing graduates’ perceived readiness for the human-AI collaborative and sustainable ethos of Industry 5.0. 2.6 Conceptual Framework The previous sections analyzed the global frameworks on 21st-century skills, AI literacy, and Industry 5.0 readiness within the disciplinary context of architectural education. 21C skills cultivated through traditional curriculum (TC_21C) and AI Literacy (AILS) form the backbone of the conceptual model as the independent variables (Khodeir & Nessim, 2020; Wang, Rau, & Yuan, 2022). AI-Enabled 21C skills (AI_21C), a novel construct, is the mediating variable (Wicks & Paulus, 2022), reflecting a hybrid integration of both human and digital capabilities; while Industry 5.0 Readiness (IR5) is the dependent variable (European Commission, 2021). Figure 1 illustrates the AIRS5.0 model, proposing how traditional curriculum and AI-tools based competencies interact through AI-enabled skills to shape graduates’ readiness for Industry 5.0. Figure 1 AIRS-5.0 conceptual model illustrating the relationship between Independent Variables (TC_21C and AILS), Mediating Variable (AI_21C) and Dependent Variable (IR5) Source: Author Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 308 The following section outlines the development and initial validation of this model through the construction of the AIRS-5.0 scale and its pilot reliability testing. 3. Methodology 3.1 Instrument Development and Validation Process 3.1.1 Conceptual Framework Basis The AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0) scale builds upon existing validated frameworks and scales into a cohesive structure to coalesce traditional curriculum, AI literacy, and Industry 5.0 readiness through AI-enabled 21st-century skills. The 21st-Century skills acquired through Traditional Curriculum (TC_21C) is derived through various existing frameworks on 21C skills (Binkley, et al., 2010; OECD, 2018). It adopts the Cognitive-Interpersonal-Intrapersonal (COG-INT-INTRA) taxonomy to assess the development of 21C skills cultivated implicitly with the traditional architecture curriculum (Salama, 2015; Khodeir & Nessim, 2020; Ng, Mari, & Lin, 2022). AI Literacy (AILS) is adapted from Wang, Rau, and Yuan (2022) and Long and Magerko (2020) models encompassing four dimensions: Awareness, usage, evaluation, and ethics. It is guided by UNESCO (2023) guidelines on integrating AI competencies in higher education. In this study, these scales are contextualized within the scope of architectural education (Basarir, 2022) AI-Enabled 21st-century skills (AI_21C) is a novel construct integrating traditional 21C competencies with AIenabled design workflows incorporating analytical and simulation-based design processes. It builds upon the AIhuman collaborative competencies developed by Wicks & Paulus (2022) and contextualized in architecture education by Salhab & Aboushi (2025) to represent hybrid human-AI competencies. It is assessed within the existing curriculum framework (NCRC - Architecture, 2025) and its sub-domains: DSW, AST, HTC, and PP. Industry 5.0 Readiness (IR5) is based upon the Work Readiness Scale (WRS) by Caballero, Walker, & FullerTyszkiewicz (2011) and the European Commission Industry 5.0 frameworks (2021). It encompasses four key domains: Human-AI collaboration and Ethics, Sustainability & Systems Thinking, Teamwork & Professionalism, and Career Readiness & Lifelong Learning. It is tailored to the architecture discipline through HEC (2025) and PCATP curricular guidelines and existing studies (Suleman, 2024). Combined, these constructs create an instrument that links traditional curriculum, AI literacy and hybrid humanAI capabilities with Industry 5.0 readiness; herein referred to as the AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0). 3.1.2 Item Generation and Scale Structure The study looks at existing frameworks in the related domains and develops the AIRS-5.0 scale according to recommended procedures for new scale design proposed by DeVellis (2017) and Boateng et al (2018). A multi-phase approach was adopted, following Fowler’s (2014) recommendation for ensuring conceptual clarity and content validity while reducing survey fatigue for a multi-construct scale. Version 1.0 (n = 60) and 2.0 (n = 53) of the instrument were exploratory drafts that were compiled from global 21C frameworks and mirrored Khodeir & Nessim (2020), connecting checklist-style statements mapping COG-INTINTRA clusters across the four curricular domains: Design Studios & Workshops (DSW), Allied Sciences & Technology (AST), History Theory & Criticism (HTC), and Professional Practice (PP). It served as the basis to identify key competencies and terminology for later Likert-scale development. The remaining three constructs were modeled as 7-point Likert-scale (1 = Not at all, while 7 = To a great extent) for quantitative assessment, with Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 309 selective criteria in each domain after discussion with peers. Preston & Colman (2000) and Joshi, Kale, Chandel & Pal (2015) argue that 7-point formats reduce central-tendency bias and offer superior balance between reliability, validity, and respondent preference. Version 3.0 (n = 106 items) also transformed TC_21C into 7-point Likert-scale. The four constructs, TC_21C, AILS, AI_21C, and IR5, were defined, and progressively refined. A sub-section for institutional support and curricular exposure was also added following UNESCO (2023) higher-education AI guidelines. Voluntary consent was added to the questionnaire at this stage as well. Versions 4.0 and 5.0 (n = 84 items) were refined on the basis of reducing survey stress and to streamline the process. However, expert feedback and a small pilot study from 14 students established that clarity and sub-domain stability had been sacrificed to reduce the item count. By this stage, the four constructs and their sub-domains had been clearly established. The final pilot instrument, version 6.0, (as seen in Annexure A), increased the number of final items to ninety-one (91), while switching the statement-based questions to more precise phrasing for better cognition. The questionnaire was structured along these lines:  Section A: Informed consent and demographics.  Section B (52 questions): TC_21C, subdivided into 13 items representing COG-INT-INTRA taxonomy correlated against four knowledge areas  Section C (15 questions): AILS, split into 11 questions for awareness, usage, evaluation, and ethics, with four (4) supplementary questions for Institutional support. Respondents were also required to report their frequency of AI tools. Those who respond with “No, never” will be redirected directly to IR5, skipping AILS and AI_21C.  Section D (12 questions): AI_21C, divided along the four knowledge areas  Section E (12 questions): IR5, divided along the four domains of human–AI collaboration & ethics (HAI), sustainability & systems thinking (SUS), teamwork & professionalism (TP), and career readiness lifelong learning (CRLL).  Three (3) qualitative questions on curriculum and AI integration were also added. The pilot testing utilized this final instrument and yielded responses from both thesis-year students (valid n = 33) and faculty (valid n = 14). The results found excellent internal consistency (Cronbach’s α = 0.83 – 0.92) with a risk of redundancy shown through relatively high Mean Inter-Item Correlation (MIIC = 0.62 – 0.83). These results form a suitable basis for large scale factorial testing (EFA/PLS-SEM) while demonstrating conceptual coherence. Overall, this six-stage development process yielded a coherent, multi-domain instrument that linked traditional curriculum, AI literacy, and Industry 5.o readiness through AI-enabled 21C competencies. It is contextualized in the architecture curriculum through a distinct connection to the core knowledge areas. The 91-item scale prioritized conceptual clarity and coverage while reducing survey fatigue (DeVellis, 2017; Boateng, Neilands, Frongillo, MelgarQuinonez, & Young, 2018). A mirrored instrument was developed for the faculty to triangulate their assessment of students’ capabilities and readiness. Same 7-point Likert scale with compact phrases for identical constructs and sub-domains were used to ensure statistical compatibility (Haynes, Richard, & Kubany, 1995; Fetters, Curry, & Creswell, 2013). 3.1.3 Expert Review and Qualitative Content Validation Fifteen reviewers from PCATP-accredited architecture programs, chosen on the basis of their expertise in design pedagogy and digital integration in architecture curriculum. They were asked to review version 5.0 (n = 84 items) for item clarity, redundancy, and construct alignment. Their feedback revealed that Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 310  Several items were ambiguous or double-barreled, potentially confusing respondents (Fowler, 2014).  Sentence-style questions increased cognitive stress, while some disciplinary wording linked to the four knowledge areas (DSW, AST, HTC, and PP) needed strengthening.  Conceptual overlap between AILS and AI_21C.  Some dub-domains only had two items, raising questions of internal validity and factorial identification (Clark & Watson, 1995). Based upon their feedback, it was decided to shift the overall language of the questionnaire from long sentencebased items towards concise phrasing to reduce ambiguity and cognitive load. Items under communication within the Intrapersonal domain of TC_21C construct were subdivided into Oral and Visual Communication, highlighting the difference within the architecture domain. This increased the TC_21C items from 48 to 52. 2 new items were added under Institutional support for AI integration (total 4 items), while AI_21C and IR5 were standardized with 3 items per sub-domain. This led to the finalized Version 6.0 with 91 Likert-scale items, which was used for pilot testing. The review group also filled and reviewed a mirrored faculty version of the final questionnaire (Version 6.0). 15 faculty members verified that item phrasing was clearer and applicable from both students and tutors’ perspective. They also verified that constructs, sub-domains, and individual items were accurately linked and represented students’ learning abilities. This multi-phase expert review and pilot survey excluded the need for statistical indices like CVI/CVR (Haynes, Richard, & Kubany, 1995; Boateng, Neilands, Frongillo, Melgar-Quinonez, & Young, 2018). 3.2 Pilot Study and Initial Validation 3.2.1 Pilot Testing and Sample Version 6 of the instrument (n = 91 items) was used for a pilot survey to test preliminary reliability and internal consistency of the finalized AIRS-5.0 scale, and to verify that the revised items were clearly interpreted and that the four constructs (TC_21C, AILS, AI_21C, IR5) and their sub-domains performed consistently prior to large-scale data collection. Lahore was chosen as the site for this pilot survey as it holds the largest concentration of Pakistan Council of Architects and Town Planners (PCATP) accredited Bachelor in Architecture (B. Arch) programs (PCATP, 2024). These ten (10) institutions include both public and private institutions as well as a representational balance of gender and students from across the country. From the students enrolled in 5-year B. Arch program, only thesis year students were included as part of the study as they had ample experience of studying through the traditional curriculum, had adequate exposure to generative AI tools, and were on the cusp of joining professional practice. The total population of B. Arch thesis year students in these ten institutions is approximately 300. Since this phase of the scale validation only aimed for item clarity and preliminary reliability rather than generalization, a pilot sample of approximately 30 students was considered adequate in line with recommendations by Johanson & Brooks (2010), Hertzog (2008), and DeVellis (2017). Data was collected with standard ethical procedures. Participation was approved by HoDs and thesis tutors. Informed consent statement was included in Section A. Any participant who selected No in response was directed to the end of the survey instrument. Data was collected anonymously and reported in aggregate. The survey was administered online via Google Form links. Average completion time was 12-15 minutes. The collected data was exported to SPSS v29 for descriptive analysis and reliability testing. 43 final-year B. Arch students from ten PCATP accredited architecture programs in Lahore participated in the survey. Standard data cleaning procedures were applied; one (1) student did not consent, seven (7) responses showing low variance (SD < 0.40, following Meade & Craig (2012) and Hair, et al. (2022)) and two (2) responses showing straight-lining (similar response > 60/91) were removed following Curran’s (2016) guidelines. 33 valid responses were retained. Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 311 Initial recruitment aimed to include both public and private institutions; however, bureaucratic procedural delays in several public-sector universities meant that final responses were heavily skewed toward private universities (30 private, 3 public). Since this pilot study aimed for reliability testing and not representative inferences and generalizations, the unequal distribution did not affect the statistical purpose of internal consistency analysis (DeVellis, 2017). Fifteen (15) tutors received the mirrored faculty version of AIRS-5.0; fourteen (14) responses were retained after one was flagged for low SD. These responses were used for triangulation and interpretive validation (Fetters, Curry, & Creswell, 2013). Preliminary results showed excellent internal consistency. Cronbach’s α was recorded between 0.83-0.92 across constructs (Tavakol & Dennick, 2011). Inter-item correlations within each construct ranged from r = 0.55-0.87, indicating strong internal consistency with risk of redundancy (DeVellis, 2017; Boateng, Neilands, Frongillo, Melgar-Quinonez, & Young, 2018; Hair, Hult, Ringle, & Sarstedt, 2022). Even though the pilot sample was small, it followed the criteria for initial scale validity and item refinement according to Johanson & Brooks (2010) and Hertzog (2008). These pilot results confirmed that the AIRS-5.0 instrument demonstrated high reliability and construct coherence. This instrument can thus justifiably be used for larger-scale validation and factor-analysis. 4. Data Analysis and Findings The pilot study (valid n = 33) was conducted to examine the reliability and internal consistency of the final questionnaire (version 6.0) of the newly developed AI Readiness and 21st-Century Skills Scale for Industry 5.0 (AIRS-5.0) instrument prior to large-scale validation. Data was collected through online Google Forms. It was later analyzed in SPSS 29 with mediation analyses conducted through PROCESS Macro Model 4 (Hayes, 2022). Student data was roughly equally distributed between male (18) and female (15) with respect to gender, and regularly (15) and occasionally (17) with respect to their self-reported usage frequency of AI tools. One student admitted to using AI tools when specified by the tutors. This suggested a wide-spread integration of AI tools within academic setting. Descriptive statistics, as shown in Table 1, showed that students rated themselves as moderately competent within all constructs (mean = 3.86 - 4.30). However, faculty rated them consistently lower in each construct (mean = 2.62 – 4.06). This indicates a consistent perception gap. The lowest scores for both set of respondents were reported for Institutional Support for AI Integration (AIIS) highlighting an academic gap. Overall, the distributions were fairly normal with skewness < ±2, and kurtosis < ±3 (Hair, Hult, Ringle, & Sarstedt, 2022). Given the unequal sample sizes (students n = 33; faculty n = 14) and the exploratory nature of the pilot study, inferential statistics such as Cohen’s d were used to report effect sizes rather than relying on significance testing. The differences were statistically significant in AIIS, AI_21C, and IR5. Table 1 Descriptive Statistics for Thesis-Year Students and Tutors showing Mean Values and Standard Deviation across Core Constructs of AIRS-5.0 Construct Mean value Std. Deviation Mean Comparison Students Tutors Students Tutors Cohen’s d TC_21C 4.30 4.06 1.31 1.20 0.30 AILS 3.99 3.31 1.58 0.80 0.42 AI_21C 4.04 2.99 1.59 0.90 0.73 IR5 3.98 3.04 1.49 1.08 0.68 AIIS 3.86 2.63 1.93 0.91 0.73 Source: Author’s work using SPSS v29 Pakistan Journal of Social Sciences, Vol. 45(4) 2025, 303-318 318 Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach's alpha. International Journal of Medical Education, 2, 53-55. doi:10.5116/ijme.4dfb.8dfd Tedjosaputro, M. A. (2019). Concerning digital design in architecture pedagogy. The 3rd International Conference on Empathic Architecture (ICEA-2019) (p. online). Surabaya: IOP Publishing. doi:10.10881/17551315/490/1/012011 UNESCO. (2023). Reimagining our futures together: A new social contract for education. Paris: United Nations Educational, Scientific and Cultural Organization (UNESCO). Récupéré sur https://unesdoc.unesco.org/ark:/48223/pf0000379707 Voogt, J., & Roblin, N. P. (2012). A comparative analysis of international frameworks for 21st century competences: Implications for national curriculum policies. Jounral of Curriculum Studies, 44(3), 299-321. doi:10.1080/00220272.2012.668938 Wang, B., Rau, P.-L. P., & Yuan, T. (2022). Measuring user competence in using artificial intelligence: validity and reliability of artificial intelligence literacy scale. Behaviour and Information Technology, online. doi:10.1080/0144929X.2022.2072768 Wicks, D., & Paulus, M. J. (2022). 21st Century Learning Skills and Artificial Intelligence. Dans SPU Works (pp. 152168). Seattle: Seattle Pacific University. Récupéré sur https://digitalcommons.spu.edu/works/217 Woo, H. (2025). Human–AI collaboration: Students’ changing perceptions of artificial intelligence in learning contexts. Sustainability, 17(18), 8387. Récupéré sur https://doi.org/10.3390/su17188387 Acknowledgments The authors are grateful for comments from two anonymous referees. Disclosure statement No potential conflict of interest was reported by the author(s). Disclaimer The views and opinions expressed in this paper are those of the authors alone and do not necessarily reflect the views of any institution. Muhammad Aarez Ali is an Assistant Professor at Razia Hassan School of Architecture at Beaconhouse National University, Lahore, Pakistan. He is also a student of MS in Project Management from Superior University, Lahore. He got a Bachelor in Architecture Design from Beaconhouse National University, Lahore. His research interests are architecture design, pedagogy, and influence of AI in the architecture industry. ORCID: 0009-0001-16118850 Malik Javied Anwar is the HoD at the Department of Emerging Management Sciences at The Superior University, Lahore, Pakistan. He got his Master’s in Accounting and Finance, M. Phil in Commerce and Finance, and PhD degree in Business Administration and Management from The Superior University, Lahore. His research focuses on Virtual Reality in Aviation Industry, Budgeting and Financial Oversight, Curriculum Design, and IndustryAcademia Linkage. His LinkedIn profile can be viewed at https://www.linkedin.com/in/dr-malik-javied-anwar-67890759/