Abstract
Virtual production (VP) is becoming central to film and television education, with universities offering degree programs, minors, tracks, electives, and short-term credentials. This review of 115 English-language sources, including 55 curricula from 49 higher education institutions (HEI), shows VP as a socially uneven, tool-weighted formation clustered in well-resourced Anglophone systems. Curricula overwhelmingly foreground real-time workflows, engine-driven pipelines, and stage operations over story development, audio design, and game-adjacent or interactive practices. The core tools include the Unreal Engine, motion-capture systems, and LED volumes, framed as prestige infrastructure rather than collective capacity. Programs emphasize employability, production-style blocks, and ‘learning by doing real jobs’, supporting industry transition but compressing experimentation, critique, and cross-cultural perspectives. Competency stacks map robust technical cores but reveal structural gaps in leadership, narrative, sound, and AI/ML literacy. The findings argue that evaluating VP education requires analyzing how programmes distribute technological and symbolic capital, organize human–machine networks, and produce learning spaces. Future research should model VP curricula as sociotechnical networks, measure AI integration maturity, test transferability, track longitudinal outcomes, map non-English ecosystems, and formalize stage pedagogy frameworks.
Keywords
Introduction
Contribution
Related work.
Challenges: Skills gaps, training needs, and standardization
The virtual production (VP) market is projected to expand by 14% by 2032 (Fortune Business Insights, 2024), primarily driven by advancements in artificial intelligence tools, including generative AI, NeRF, Gaussian Splatting, and hybrid workflows (Moore, 2024). However, the absence of standardized workflows poses a significant barrier to effective implementation (Digital Catapult, 2023c; Willment et al., 2024b). The BFI Skills Review (2022) has identified skill deficits in VP as critical bottlenecks and has recommended the implementation of structured training programs. Also, the VP Skills Mandala highlights deficiencies in areas such as game-engine operations, asset optimization, and supervision (Bennett et al., 2021; Heath et al., 2023). Thus, a lack of fluency in VP tools among creatives leads to mismatched expectations, while inconsistent terminology between film and technology professionals impedes collaboration (Raimundo et al., 2024; Willment et al., 2024b). Initiatives such as Disguise’s VP Accelerator and MARS Academy are promoting alignment in workflows (Foley, 2024), while the Virtual Production Skills Network (VPSN) is integrating VP into media education (Haggis-Burridge et al., 2022). Therefore, experts underscore the importance of training within live VP ecosystems (Barnett et al., 2024; Digital Catapult, 2023c). Without sufficient workforce training, the potential for growth may stagnate, as evidenced by the prevalence of unpaid internships (ResearchNester, 2024; Willment et al., 2024a).
In this context, this study asks:
Article structure
This review follows a systematic progression, establishing historical context, terminology, roles, core skills, and technological influences on VP workflows. The Methodology outlines data preparation, search strategies, analysis, sampling, and literature distribution. The analysis addresses: (1) curricular integration of VP in higher education, (2) key educational content for film programs, (3) pedagogical theories and strategies, (4) pedagogical outcomes and career pathways, and (5) emergent technologies. The results were interpreted through the theoretical lenses: Lefebvre’s production of space, Bourdieu’s concepts of field and capital, and Latour’s actor-networks. The article presents Findings and Conclusions, noting limitations and future directions.
Research background
Historical setting
Virtual production (VP) encompasses a range of techniques, extending from early compositing, such as double exposure and chroma keying, to the real-time integration of live-action and digital environments. Chroma key innovations, including the colour-difference travelling matte (Vlahos, 1964), have become the standard for digital background replacement (Fry and Fourzon, 1977; Pires et al., 2022). Today, real-time engines allow on-set compositing and iterative creative feedback (Chanpum, 2023; Chen and Kadner, 2021; Hunt et al., 2018; Kadner, 2021; Silva Jasaui et al., 2024). VP is categorized into fully animated and live-action approaches, spanning visualization, hybrid green screen, LED in-camera setups, and motion capture, all unified by real-time engines (Pires et al., 2022).
Hybrid green screens and LED walls enable natural interaction with virtual environments. Since Avatar’s Simulcam (2009), companies like Moving Picture Company (MPC) and Unity/Unreal have developed VP pipelines (Genesis) for The Jungle Book (2016) and The Lion King (2019) (Giordana et al., 2018; Tovell and Williams, 2018), advancing what Favreau described as ‘movie gamification’ (Favreau as cited in Li et al., 2022: 225). LED volumes, used in The Mandalorian (2019), improve lighting realism and actor immersion, and resolve green-screen reflection issues (Seymour, 2020). VP also supports performance capture across live action and animation (Fleischer, 1915; Menache, 2000).
Terms and definitions
‘Virtual Production’ is an umbrella concept that encapsulates the use of technology to connect the digital and physical worlds in real-time film production. This enables filmmakers to interact with the digital process in the same way they interact with live-action production. Examples of virtual production include world capture (location/set scanning and digitization), visualization (previs, techvis, postvis), performance capture (mocap, volumetric capture), simulcam (on-set visualization), and in-camera visual effects (ICVFX). The key to the successful use of this technique is to choose the right tools to solve production problems and empower the creators without detracting or distracting the crew from the content creation process, as shown in Figure 1 (Virtual Production Glossary - Virtual Production, 2024). Therefore, traditional film production has shifted post-production processes to pre-production, as shown in Figure 2. Example of virtual production workflow adapted from (Kadner, 2021: 148). Traditional versus virtual production for film (Kadner, 2019: 6).
Considering the growing demand for visual effects and the sustainable needs of the film industry, production must address carbon emission reductions from crew travel, car chases, fuel costs, and materials (Bigger Picture Research, 2020; Helzle et al., 2022; Oey, 2024). Virtual Production offers a cleaner, faster, and cheaper way to achieve spectacular effects for cinema audiences. Even so, due to the costs of energy, hardware, and data infrastructure, VP is not automatically environmentally neutral but is seen as a good step in this direction.
Virtual production (VP) encompasses computer-aided production and visualization methods that merge the physical and digital domains (Kadner, 2019; Kavakli and Cremona, 2022). Its adoption requires new departments, roles, and workflows, demanding institutional adaptation in film production structures (An, 2022b; Jorge, 2024a; Willment et al., 2023).
The consensus emphasizes that VP must integrate with cinematography, art, production design, and VFX, fostering interdisciplinary collaboration (Swiatek, 2024; ScreenSkills, 2023; Spillers and Kadner, 2021). Industry professionals note that VPs’ iterative, sequence-based workflows replace the linearity of traditional pipelines, supporting greater flexibility, as shown in Figure 2 (Grossmann et al., 2024; Silva Jasaui et al., 2024).
VP enables, therefore, earlier collaborative creative decision-making, reducing reliance on post-production (Kadner and Kadner, 2019; Xue et al., 2019). Thus, disrupting the linear workflow, transforming it into a series of interactive agile cycles in order to deconstruct hyper complexity, as Figure 2 shows.
Emergent technologies in Virtual Production workflows
Emerging technologies, such as artificial intelligence (AI) and machine learning (ML), have substantially influenced VP, facilitating procedural content generation, digital asset management, and high-fidelity real-time rendering techniques, including Deep Learning Super Sampling (DLSS), NeRFs, and AI-enhanced texture synthesis (Champagne, 2024; Grossmann, 2024; Moore, 2024; Silva Jasaui et al., 2024). These innovations extend into markerless motion capture, facial tracking, and automated animation workflows, raising ethical, workforce, and educational considerations and highlighting the need for AI literacy and cross-disciplinary training (Digital Catapult, 2023a, 2023b; Foley, 2024; Silva Jasaui et al., 2024; SkillsFuture Singapore et al., 2024; Swords and Willment, 2024).
Planning
Theoretical background
In the initial phase, the systematic search process followed the three-phase protocol outlined by Kitchenham and Charters (2007): Planning, Conducting, and Reporting. Due to the limited availability of scholarly papers on this subject, the snowballing method was also used to identify additional literature, in accordance with Wohlin’s (2014) guidelines. However, given the scarcity of academic sources, a multi-local literature review (MLR) was necessary in the second phase. Unlike traditional Systematic Literature Reviews (SLRs) (Kitchenham and Charters, 2007), which rely solely on academic articles (White or Formal literature (WL)), Multivocal Literature Reviews (MLRs) can incorporate Grey Literature (GL) (Garousi et al., 2019).
Grey Literature, often produced by practitioners such as private industry, government bodies, and independent researchers, exists outside commercial publishing and peer review, and typically does not appear in scientific journals or books (University College London, 2025). As virtual production is still in its early stages, incorporating practitioner-generated literature is essential for gaining deeper insights into industry practices. By integrating this content, one can synthesize both academic findings and practical advancements to achieve a more comprehensive, state-of-the-art perspective.
PICOC breakdown
PICOC framework applied to virtual production within higher education institutions.
Research questions
Considering these challenges, a systematic Multivocal Literature Review (MLR) of the pedagogical virtual production landscape was conducted to address the following research questions: RQ 1 How are higher education institutions currently integrating virtual production into film curricula, and how is virtual production structured within these academic programs? RQ 2 What specific tools and subject areas related to virtual production are currently emphasized within higher-education film programs, and what are the literature recommendations? RQ 3 What pedagogical theories and strategies (including industry-academia collaboration) are employed in virtual production instruction in film studies? What findings have been documented in the existing literature? RQ 4 What learning objectives and career outcomes are targeted by higher-education programs that incorporate virtual production? RQ 5 To what degree are emerging technologies (AI, machine learning) integrated into virtual production curricula in film education programs?
Search strategy
In the first stage, the search string was determined by deriving relevant keywords from the following research question: ‘virtual production’ AND (‘film studies’ OR ‘media production’ OR ‘screen industries’) AND (‘education*’ OR ‘programs’ OR ‘curricul*’ OR ‘skills’ OR ‘competenc*’ OR ‘teaching’ OR ‘artificial intelligence’ OR AI OR ‘machine learning’ OR ‘emergent workflows’ OR ‘emergent technologies’)
The review searched ACM, Scopus, IEEE Xplore, ScienceDirect, SpringerLink, Wiley, EBSCO, and JSTOR, with the results screened against the criteria, and Snowballing (Wohlin, 2014) supplemented the limited sources (Appendix A). Studies were included if they detailed VP teaching through curriculum design, instructional strategies, or assessments (Bell, 2010), or addressed VP skills/work readiness with clear links to professional competencies (Caballero et al., 2011; Caballero and Walker, 2010; Nimkulrat, 2007). General media or workforce studies without a VP context were excluded from the review. Furthermore, the exclusion of Virtual Reality (VR) and Augmented Reality (AR) filmmaking was adopted as a methodological boundary to focus specifically on traditional filmmaking contexts, rather than to imply a theoretical or comparative analysis of these emerging filmmaking forms.
For the second part of the MLR, we established another protocol to filter and extract the GL using the quality assessment guidelines suggested by Garousi et al. (2019) to identify all relevant grey literature. The source collection process has been automated using Python code. For brevity, all details are available online (see Appendix A for more information). Figure 3 shows the complete selection process. Diagram selection results, white and grey literature.
Conducting
Data analysis
The literature data extraction process followed Vaismoradi et al. (2016) framework for qualitative content and thematic analysis. The method comprises four phases: initialization, construction, rectification, and finalization, which involve reading, coding, comparing codes, defining themes, and developing narratives for research questions (Vaismoradi et al., 2016). Excerpts were extracted, coded, and themed per review question, with detailed procedures provided in the Supplementary Codebook (Appendix A).
For VP curricula data extraction, experts validated the extraction categories (Appendix A). The respondent cohort included professionals aged in their late 30s to mid-50s, with 5-35+ years of experience in roles such as VFX Compositor, VP Specialist, Professor, and Academic Researcher, representing both industry and academia. Participants from Germany, Ireland, Finland, Malta, and Portugal held qualifications from diplomas to doctoral degrees in film, immersive media, and academia. The consensus exceeded 80%, above the typical 75% threshold for Delphi studies (Diamond et al., 2014). Validation occurred before final analysis (Appendix A).
Strengths of the evidence
The evidence is robust, using systematic data extraction and clear selection criteria for English-language VP references. This enables comparisons across regions, program types, and integration levels. Geographic distribution graphs show adoption patterns.
Technologies and subject areas were clearly identified. Counting tools, such as Unreal Engine, LED walls, and MoCap, prevent bias while aligning with industry standards. Literature references contextualize curricular adoption.
Curricular analysis is based on educational theory. Pedagogical approaches are linked to Freire’s critical pedagogy, Vygotsky’s scaffolding, and Schön’s reflective practice, providing a strong academic foundation for practical VP teaching.
Categorization into competency areas enables program comparisons. Curricular emphases align with industry demands, demonstrating practical utility. The examples demonstrate institutional differentiation and career outcomes, showing current VP curricula.
The analysis highlights the integration of AI alongside literature trends. Cases like Runway AI provide clarity, while references support deeper adoption of AI/ML curricula that align with industry innovations.
Weaknesses of the evidence
This review relies on publicly available curriculum descriptions, which may omit institutions with implicit VP integrations. English-language sources introduce bias, limiting generalizability and underrepresenting programs in countries like India, China, France, Germany, and Spain, where local curricula may differ. Using the term ‘virtual production’ alone may have excluded programs that use different terminology.
The lack of qualitative data limits insights into teaching methods, outcomes, and feedback. Institutional differences and the gap between curriculum intent and student experience remain unexplored. Tool frequencies lack context, leaving emerging technologies underexplored. Adoption levels may reflect resources rather than pedagogical effectiveness.
AI/ML adoption is likely underreported, as programs often omit embedded VP tool usage. Integration patterns may reflect institutional resources, faculty expertise, or proximity to an industry hub. These findings should be considered indicative rather than definitive.
Limited qualitative data on student outcomes, faculty perspectives, and assessment practices constrain pedagogical evaluation. The sustainability of industry-integrated methods remains uncertain due to expertise and resource requirements.
Findings
This section presents key findings, structured to separate educational programs (55) from other grey literature, enabling a clear analysis of curricula, literature, and observed patterns.
RQ 1: How are higher-education institutions currently integrating virtual production into film curricula, and how is virtual production structured within these academic programs?
Data were extracted from 49 higher-education institutions to assess VP techniques in film studies. Curricula mentioning ‘Virtual Production’ in English were analyzed (Appendix A) according to the defined categories in the data analysis section. This enabled a comparative analysis across regions and provided a reliable map of VP education.
As illustrated in Figure 4, the strong concentration of programs lies in the USA and the UK, with 29 programs, approximately 59% of the programs in the study universe. Together with Australia (6), Canada (5), and Ireland (3), approximately 88% of the programs are in Anglophone countries. The rest of the world, represented by China (1), India (1), Switzerland (1), the Netherlands (1), Germany (1) and Sweden (1) represents approximately 12% of the analysis sample. Distribution graph: Higher Education Institutions (HEI) geographic location.
Europe (excluding UK and Ireland) comprises a total of 4 programs, roughly equal or less than a single country like Australia (6), showing a tendency towards specific hubs rather than the continent as a whole. In the same line of thought, Asia (India and China), with 2 programs and 0 entries from Africa, Latin America, and the Middle East, exhibits that the ‘global map’ of VP Education is actually a global-North (English-Language map), shaped by the search criterion ‘curricula in English’.
Since approximately 90% of the identified VP programs are in a small set of Anglophone countries, access to formal VP Education in English is geographically uneven. Therefore, students and industry in underrepresented regions may depend more on online training, Educational and industry partnerships or study-abroad institutional agreements to gain access to VP Workflows.
Curriculum integration of virtual production in academic programs
Fully integrated virtual production degree programs
Figure 5 shows that Bachelor´s (General) is the most common format, with 10 entries. Undergraduate or short-cycle programs (Bachelor´s, Minors, Micro-credentials, Certificates, Diplomas) encompass almost 70% of all programs, leaving roughly 30% for Postgraduate and Professional programs (MA, MSc, MFA, AAS, MPS, graduated Diplomas and certificates). Distribution graph: Higher Education Institutions degrees.
The 6 Micro-credentials show that HEIs are also using VP as modular or stackable learning, not just for full degrees. Furthermore, the various flavours of bachelor’s, master’s, diploma, and certificate programmes suggest possible local branding and accreditation logic, even when the curricular content is likely similar.
From a broader perspective, VP Education is currently being designed primarily with a foundational training and upskilling focus. There is the ecosystem of specialized VP masters or research degrees is yet not wide. In this sense, advanced expertise will likely still be developed through on-the-job practice and a small number of graduate programs.
VP specialization track integration within academic degrees
Figure 6 depicts the percentage of VP specialization tracks within various academic degrees. Advanced degrees and micro-credentials show a clear positive relation between the degree level and VP specialization. This means that tiny, short programs have a narrow focus in VP, as do full postgraduate degrees. Percentage VP specialisation within academic degrees.
On the other hand, the Bachelor’s (BSc general (0.19) versus BA/Bachelor general (0.47)) degrees are science-labelled film/media programs that are less VP-focused than more ‘arts’ labelled ones. Furthermore, some Master’s programs (e.g. MSc at 0.42, MFA at 0.75) break the 100% pattern. Therefore, not all postgraduate degrees guarantee a deep focus on VP.
Based on this analysis, many HEIs currently treat VP mainly as advanced specialization/upskilling, not as the default core of broad undergraduate study.
RQ 2: What specific tools and subject areas related to virtual production are currently emphasized within higher-education film programs, and what are the literature recommendations?
The analysis identified the following tools and subject areas: Each institution was incorporated once per tool and subject area to circumvent duplication, thus avoiding duplication due to multiple programs within the same university. The results are shown in Figures 7 and 8. Distribution graph of tools taught in universities. Distribution graph of subject areas taught in universities.
The graphic shows key patterns in real-time/virtual production tools (Unreal, Mocap, LED Wall, VR/AR, real-time rendering compositing tools, XR platforms, such as Disguise, and tracking tools), approximately 81%.
Comparing the number of mentions related to Digital Content Creation (DCC) apps (Maya, Houdini, ZBrush, Blender) (12) with Unreal Engine (40) the graphs show a game-engine centricity. On the other hand, the underweight audio tools, despite sound’s importance, the tools and workflows discussed are image- and stage-centric.
In this line of thought, if the real-time tools share is approximately 80%, then training, Research & Development, and Budget allocation should tilt heavily toward real-time, engine-based virtual production, with audio, cameras, and traditional post treated as integration points around that core.
Subject areas
Figure 8 shows the key subject areas taught, along with the number of distinct programs teaching each.
The top five subjects, with approximately 67% of all mentions, are present in the analyzed programs, indicating a clear focus on real-time and on-set technical topics.
On the other hand, scriptwriting (3) and Previz & Storyboard are marginally represented. Game Design & Development (3) appears with low frequency, although VP is acknowledged as game-engine adjacent. Also, Audio and Sound Design (3) is again underrepresented compared to image-centric topics, echoing the tool chart observations on this same topic.
Considering the strong focus on on-set tools practice and weak focus on story and audio, graduates may become excellent stage operators or technologists, but less trained to manage end-to-end cinematic storytelling with VP, unless these modules are combined with separate writing and post-production courses.
Literature recommendations on tools and teaching subject areas
Film schools and media programs are increasingly integrating virtual production (VP) technologies into curricula to meet industry demands for hybrid skills. Curricula are now designed around real-time engines, volumetric capture, LED wall production, and AI-driven workflows (Calawerts et al., 2024; Digital Catapult, 2023a; Dooley and Emery, 2023). These tools support interdisciplinary education, merging cinematography with computational techniques.
Real-time game engines and virtual environments
Game engines, particularly Unreal Engine and Unity, are recognized as foundational for VP, covering previsualization, real-time rendering, and final pixel outputs (An, 2022a; Digital Catapult, 2023a). Courses that embed game-engine training enable students to manipulate 3D environments, test workflows remotely, and apply interactive simulations (Jushchyshyn and Parks, 2021). Scholars argue that this convergence of film and game production demands hybrid expertise, requiring filmmakers to also grasp core game design (Boutellier and Raptis, 2023; Kavakli and Cremona, 2022).
Motion capture and volumetric capture
Motion capture (MoCap) is central to VP education as it connects theory with practice in performance animation. Literature recommends courses covering full-body capture, facial tracking, and markerless systems (Digital Catapult, 2023a; Hendry, 2024; Pohl, 2024a). For example, Flinders University’s ‘The Void’ studio integrates Vicon MoCap systems for both teaching and R&D, allowing students to direct performers and process movement data within real-time engines (Somerville, 2024).
Previsualization and virtual cinematography
Previs techniques help students plan complex productions digitally. Literature suggests modules on Techvis and digital storyboarding (Pohl, 2024a; Xue et al., 2019), virtual location scouting with VR replicas (Jorge, 2024b; Materska-Samek et al., 2023), virtual pre-lighting for simulated illumination (Swiatek, 2024), and cineDESK camera simulation platforms for immersive cinematography workshops (Hendry et al., 2023). Augmented reality (AR) expands these workflows by allowing integration of CGI assets into live rehearsal spaces (Xue et al., 2019).
LED walls and virtual set technology
LED volumes enable students to simulate in-camera VFX, blending physical props with real-time environments. Studies recommend hands-on training in synchronizing LED panels with camera tracking and lighting (Haggis-Burridge et al., 2022; Venckute and Jablosnkyte, 2024). Flinders University’s Void Studio exemplifies this approach, teaching LED configuration and stage operation (Barnett et al., 2024; Jason Bevan, 2024).
Extended reality (XR) and AI
XR tools and AI-driven methods enhance immersion and rehearsal. AR/VR let students engage directly with digital environments, merging physical performance with virtual assets (Cannavò et al., 2023; Xue et al., 2019). This supports more embodied, creative, and technologically adaptive training.
RQ 3: What pedagogical theories and strategies (including industry-academia collaboration) are employed in virtual production instruction in film studies? What findings have been documented in the existing literature?
Based on the extracted information from the analyzed curricula (Appendix A), VP education was analyzed across undergraduate, postgraduate, and certificate programs, linking teaching strategies to pedagogical frameworks and industry practice. The examples provided illustrate specific course focuses that suggest guiding principles without compromising the completeness of their curricula.
Undergraduate programs: Hands-on and integrated learning
In general, undergraduate programs use project-based learning and practice-based studio work with a focus on storytelling and reflection as assessment. For example:
Project-Based Learning: At Breda University, VP is integrated through team projects and industry input (Breda University of Applied Sciences, 2024), while Chapman employs workshops (Chapman University, 2024), and York focuses on storytelling (York University, 2024).
Practice-Based Learning: Ahalia University’s BVA program merges theory, studio practice, and on-the-job training (Ahalia School of Media Studies and Future Technologies (ASOMSAFT), 2024). The BA in Immersive Media at IADT, BIFE, and Ballyfermot College combines storytelling, 3D modelling, and immersive media production (Ballyfermot/IADT/BIFE, 2024). Similarly, City of Liverpool College’s BA integrates theory with world-building and 3D animation (City of Liverpool College University Centre, 2024). In contrast, Queensland University imparts VP concepts through in-camera VFX (Queensland University of Technology, 2024).
Assessment combines practice and reflection: Lynn University mandates a four-course thesis (Lynn University, 2024), Savannah College of Art and Design (SCAD) requires thesis films (Savannah College of Art and Design, 2024), while portfolios/blogs facilitate critical reflection (City of Liverpool College University Centre, 2024).
Postgraduate programs: Advanced practice and research integration
Overall, the Postgraduate VP programs focus on specialization through practice-led, research-driven models, aligned with a reflective practitioner, promoting interdisciplinary learning and internships or industry projects. For example:
Georgia State University’s MFA program seamlessly integrates cinema theory, production training, internships, and access to LED stages (Georgia State University, 2024). At Texas A&M University, students engage in extensive collaborative projects (Texas A&M University, 2024).
Filmakademie Baden-Württemberg combines project work with research mentorship (Animationsinstitut, 2024), while Solent University focuses on collaborative problem-solving (Solent University, 2024). Nottingham Trent University merges research with work-based learning (Nottingham Trent University, 2024). In alignment with Schön’s reflective practitioner model (Schön, 1983), Flinders University emphasizes reflective practice (Flinders University, 2024c, 2024d). Industry integration is a pivotal aspect, with internships and partnerships being central to the programs at Georgia State, Nottingham Trent, and De Montfort (De Montfort University, 2024; Georgia State University, 2024).
Innovative delivery formats, such as block teaching (De Montfort University, 2024) and hybrid postgraduate certificates at the National Film and Television School (NFTS) (National Film and Television School (NFTS), 2024), cater to a diverse range of learners. The interdisciplinary structures at Solent, York, and Zurich universities reflect professional collaboration (York University, 2024; Zurich University of the Arts (ZHdK), 2024). Assessment methods combine practical and research outputs. Students at Nottingham Ningbo produce VP works accompanied by a written exegesis (University of Nottingham Ningbo China, 2024), while Greenwich requires ‘illustrated reports’ to complement creative projects (University of Greenwich, 2024).
Certificate and diploma programs: Focused skills and industry alignment
In overview, certificates and diplomas focus on employability and offer compressed VP skill training, work-based learning, strong studio partnerships, and capstone portfolios as assessments. For instance:
Humber College’s postgraduate certificate focuses on VP pipeline skills through capstone projects (Humber College, 2024). The Regional VP Academy combines internships with project-based studio work (Regional Virtual Production Academy - Bay Area Center for Creative Careers (BACCC), 2024a, 2024b, 2024c, 2024d). Similarly, Flinders University designs curricula around applied industry projects (Flinders University, 2024a, 2024b, 2024c, 2024d), while Queensland University offers students opportunities to showcase their work in real-world productions (Queensland University of Technology, 2024).
Collaboration with major studios enhances training: Sheridan College partners with Pinewood Toronto Studios (Sheridan College, 2024a, 2024b, 2024c), and SCAD collaborates with Netflix and HBO (Savannah College of Art and Design, 2024). Likewise, NFTS provides exposure through professional partnerships (National Film and Television School (NFTS), 2024).
Institutions emphasize practical immersion: Griffith University and Hunter College focus on internships and professional collaborations (Griffith University, 2024; Humber College, 2024). In Europe, the Zurich University of the Arts (ZHdK) and the University of Greenwich align with industry through real-world studios and professional networks (University of Greenwich, 2024; Zurich University of the Arts (ZHdK), 2024). Blended models combine online and in-person training methods. IADT offers a Certificate in Real-Time VP (Institute of Art Design + Technology (IADT), 2024), while Sheridan College integrates online seminars with workshops at Stage 10 (Sheridan College, 2024b, 2024c).
Partnerships with studios enhance employability: Vancouver Film School collaborates with Pixomondo (Vancouver Film School, 2024), while Kwantlen Polytechnic incorporates demo-reel production (Kwantlen Polytechnic University, 2024).
The assessment prioritizes practical deliverables – projects, portfolios, and demo reels – that reinforce professional readiness (Sheridan College, 2024b, 2024c).
Pedagogical theories/strategies based on VP curricula data extraction.
Based on this analysis, a common logic core emerges centred on authentic, production-like projects. It is rebranded through multiple frameworks (PBL, Work-based, Problem-Based Learning, OJT), suggesting a theorized convergence on ‘learning by doing real jobs’. Regarding critical reflection (exegesis, illustrated reports, reflective practice), although more explicit at postgraduate levels, it also appears in highly vocational contexts (certificates, diplomas). This evidence, in contrast, shows that even employability-driven courses are being ‘academised’. Furthermore, block teaching, intensive workshops, and demo-reel assessments mimic production cycles and deliverables, suggesting that time is structured more like a production film shoot than a traditional semester.
Given this context, and because VP pedagogy currently resides at the crossroads of practice, research, and industry, curriculum designers must carefully balance industry-driven, project-intensive models with robust critical and research components. This approach helps avoid the risk of training professionals solely in industry-specific skills and tools, while also safeguarding the independent knowledge of VP production.
Findings in the existing literature
Pedagogical theories/strategies based on literature findings.
Student-centred pedagogies
Boutellier and Raptis (2023) emphasized co-created learning design to foster student ownership, aligning with Rossi’s (2023) integrative model of inclusive learning (Values, Context, Content, Assessment, Evaluation). Drawing on Freire’s notion of students as ‘critical co-investigators’, learners engage in inquiry rather than passive reception (Boutellier and Raptis, 2023: 72; Freire, 2014: 81). In VP curricula, instructors guide students through Vygotsky’s zone of proximal development (Vygotskiĭ, 1978). Realistic 3D environments allow contextualized practice with industry tools and support mastery through interactive experiences (Boutellier and Raptis, 2023). Csikszentmihályi’s concept of ‘flow’ suggests that VP can enhance focus and enjoyment in learning (Csikszentmihalyi, 1990; Flow Theory - an overview | Science Direct Topics, 2025).
Immersive and engaging learning environments
Motion capture (MoCap) bridges conceptual learning in VFX, animation, XR, and game design with practice, helping students to internalize complex concepts (Najafi et al., 2024). Educators also integrate theories such as Rudolf Laban’s movement analysis into MoCap training, as at Flinders University’s stage, The Void (Musolino, 2024). Augmented reality (AR) addresses the challenges faced by acting students trained using naturalistic methods (Cannavò et al., 2023).
Collaborative and interdisciplinary learning
VP requires collaboration across directing, cinematography, animation and game development. Success depends on explicit teamwork training (Dooley and Sexton-Finck, 2017). Strategies include group contracts, communication exercises, and reflection on team dynamics (Sabal, 2009). At Filmakademie Baden-Württemberg, interdisciplinary projects mirror industry workflows, enabling real-time collaboration among specialists (Bennett et al., 2023; Helzle et al., 2022; Murray et al., 2019).
Learning-by-doing and iterative practice
VP pedagogy stresses experiential learning: students ‘jump in and figure it out’ (The Animation Workshop/VIA University College, 2022: 32). The ideal method combines observation, supervised practice, and mentorship, as in the Unreal Fellowship by Epic Games (Pohl, 2024b). Agile methodology is applied to project-based learning, where capstone projects assign students creative roles, while faculty act as product owners working with industry clients (Jorge and Kadner, 2019).
Blended and adaptive delivery modes
Blended learning combines online preparation with in-studio practices. Flipped classrooms allow students to study software basics remotely and then apply their skills in supervised collaborative sessions (Henry and Maric, 2023).
Industry integration and skills development
Rapid VP evolution requires collaboration between academia and industry. Partnerships grant access to equipment and training while supporting workforce development (Turku University of Applied Sciences, 2024). Curricula integrate domains such as lighting, cinematography, and digital imaging (Digital Catapult, 2023a: 35). Industry leaders stress that storytelling fundamentals remain central despite technological advances (Kadner, 2021: 103). New competencies include programming, real-time 3D, and game-engine operation, with lifelong learning being essential as tools evolve (Kadner, 2021: 103).
Diversity and inclusion initiatives
Equity in VP education requires outreach and mentorship for underrepresented groups, particularly women and gender-diverse students (Willment et al., 2023). Early exposure counters biases (Lauzen, 2025), and visible role models encourage participation. Inclusive practices in team dynamics and access support equitable training (Thomas and May, 2010). Diverse pipelines from the film, gaming, and STEM fields strengthen the industry (Willment et al., 2024b).
The scope of this analysis highlights the application of a pedagogy that is student-centred, scaffolded, situated/immersive, experimental/iterative, collaborative, and interdisciplinary. Also, favouring primarily adaptive teaching, blended/flipped learning, industry partnerships, mentorship modes grounded in fundamentals and self-learning, and diversity and inclusion initiatives.
There is a clean theory line-up: scaffolding, ZDP and experimental/interactive, agile methods, immersive tech, and flow theory. Therefore, VP could become a convergence point for classic learning theory, software engineering, and performance studies. Diversity and inclusion are framed not only as ethical principles but also as an educational strategy, tying social justice directly to labour-market needs. The emphasis on fundamentals and self-directed learning shows awareness that tools change quickly, so pedagogy could target meta-skills (story, design, learning-to-learn) more than any given software.
Operating at the intersection of immersive tech, industry workflows, and rich pedagogy (Freire, Vygotsky, Flow, agile), VP could serve as a prototype for future media/film curricula more generally. However, this has a direct implication for the institution’s investment in staff development, so teachers can actually orchestrate this complex mix rather than default to tool demos.
RQ 4: Which learning objectives and career outcomes are targeted by higher-education programs that incorporate virtual production?
Learning outcomes and competencies
Despite the variety in program structures, there is substantial convergence in learning outcomes across the 49 analyzed virtual production university curricula (Appendix A).
Common learning objectives and competencies
Common learning objectives grouped into several broad competency areas.
Program focus and specializations.
Unique program focuses and specializations
Regarding common learning objectives, the evidence shows several competency stacks: 1. Core tech pipeline combining VP workflows, immersive tech, 3D VFX, Mocap, and virtual camera); 2. Creative Layers (Story and Content direction); 3 Professional Layers (Collaboration & Project Management and Emerging Topics).
On the other Hand, considering the Program’s focus and specializations, several axes are exposed, with a broader scope in Technical Engineering, creative filmmaking, VAD, Environment and Stage operations, Game and interactive media, and research and pedagogical. Thus, Table 6 exhibits four tech-heavy competencies versus one creative competency (storytelling). This suggests that the VP curriculum thematizes Story but structurally privileges pipeline fluency. The presence of emerging topics & Professional practices, Research & Professional Certification, and Learning objectives and competencies means some VP programs are explicitly positioned as innovation, not just workforce training.
In this mind, program design can be modelled based on the common competency core (Table 5) and specialization modules (Table 6). So institutions extending or benchmarking degrees should first guarantee the share core (engines, immersive tech, 3D, MoCap, story, collab) and then ‘plug-in’ one or more of the listed specialization profiles rather than investing everything from scratch.
Career outcomes
Career outcomes.
From this analysis emerge 3 clusters: 1. Technology/Pipeline (Unreal generalist/specialist/technichian; VP tech artist/technician, VFX/VP pipeline developer, Lighting, mocap); 2. Creative (3D/VFX Artist, animator, concept artist, director, Director of Photography, editor, game designer, immersive media producer); 3. Managerial (producer, pipeline tech director, supervisor, content manager). Also, the Unreal/engine-centric titles are over-specified (generalist, specialist, on-set tech, VP tech artist, VP technician), showing the existence of an engine-specific, separated labour market.
From another perspective, audio (1 role) remains marginal on the image-centric programs. Moreover, several jobs are transmedia (game designer, social/online content, immersive media producer), suggesting VP education is already leaking beyond film/TV into wider screen industries. Furthermore, the existence of both pipeline developers and pipeline directors/supervisors implies a maturing hierarchy where VP isn´t just tools but an organizational structure.
This means that VP programs should guarantee all three bands to avoid training students who can operate tools but cannot either create compelling content or step into coordination or supervisory roles in VP teams.
RQ 5: To what degree are emerging technologies (AI and machine learning) integrated into virtual production curricula in film education programs?
According to the curricula (Appendix A), the integration of AI/ML in VP education is rare. Only four programs (UK, Ireland, Canada, USA) mention AI, while 50 of 54 omit it. Mentions range from full integration to isolated modules, but the depth of these features (e.g. AI-driven animation tools) is unclear.
Explicit AI/ML integration in virtual production programs
Of the surveyed programs, only a small subset mentioned AI or ML integration in their virtual production curriculum. Four institutions highlighted the importance of AI integration: The Academy of Contemporary Music (ACM), affiliated with Middlesex University in the UK, extensively incorporates AI technology, particularly emphasizing ‘AI-driven animation tools’ (Academy of Contemporary Music and Middlesex University, 2024). This suggests a structured embedding of AI functionalities, such as animation and real-time graphics, within the educational approach. In Ireland, Ballyfermot College of Further Education, IADT, and BIFE collaborate on the BA (Hons) in Immersive Media Production, explicitly integrating ‘generative AI into specific coursework modules’ (Ballyfermot/IADT/BIFE, 2024). These modules likely emphasize AI applications in content creation and procedural generation. Humber College in Canada highlights AI through specialized expertise provided by faculty, particularly in ‘AI-driven avatars and scanning digital doubles’ (Humber College, 2024). While not necessarily offering standalone courses on AI, the college employs practical tool-based AI applications embedded within virtual production instruction. At the NYU Tisch School of the Arts in the USA, AI tools such as Runway AI are thoroughly ‘integrated across various stages of the virtual production process, including concepting, previsualization, and post-production’ (NYU Tisch School of the Arts, 2024). This comprehensive integration ensures that students develop expertise with advanced AI-driven workflows, effectively positioning them in professional contexts.
Literature recommendations on emerging technologies (AI, machine learning)
Due to the limited information available in the analyzed educational curricula (Appendix A) on this topic, insights were primarily drawn from the literature.
Emergence of AI/ML in virtual production
AI and ML are permeating virtual production processes, pressuring educational programs to incorporate these topics. Recent developments show generative AI tools creating virtual production content. Image-generating models like DALL-E, Midjourney, and Stable Diffusion produce digital background art for VP shoots (Willment et al., 2023; Willment et al., 2024b). Kadner (2022) noted that AI-based ML systems automate post-production work and accelerate rotoscopy, reducing manual effort (Kadner, 2022; Willment et al., 2023). Real-time engines may predict character movements using AI, expediting animation workflows (Willment et al., 2023).
AI-assisted tools enhance performance capture in virtual production. Motion capture combines manual training and ML models to improve virtual character control (Bergeron and Kadner, 2021). By 2019, VP innovators envisioned natural language AI interfaces, with Grossmann proposing AI voice recognition for scene creation (Kadner, 2019: 49). NVIDIA uses deep learning to accelerate graphics workflows (Rick Champagne, 2024). These AI developments in VP indicate film education should address these technologies.
Integration into film education curricula
Educational institutions are responding to AI in virtual production, though curriculum integration is still in its early stages. VP’s constituent technologies, including AI/ML, are ‘increasingly assimilated into tertiary teaching programs’, this uptake still lacks a shared pedagogical framework (Barnett et al., 2024: 2). The reviewed curricula (in Appendix A) foreground Unreal Engine, VP workflows, cinematography, VFX, and real-time rendering, yet rarely specify AI/ML methods as learning outcomes or dedicated modules. This gap is partly structural: rapid software updates and the growing presence if AI-enable toolchains make it difficult for film schools to stabilise syllabi and maintain staff and infrastructure readiness. (Venckute and Jablosnkyte, 2024). Nonetheless, emerging educational projects shor how AI can be operationalised in VP learning context - for example, students used voice commands to control a virtual environment via an AI agent that modified visuals and sound, featuring an AI avatar for participant interaction (Woolbright, 2024).
Film education programs recognize the need to teach emerging technologies alongside traditional skills. Industry reports urge educational institutions to prepare students for AI/ML in production (The Animation Workshop/VIA University College, 2022). A 2022 analysis emphasizes understanding AI/ML implications in media production and questions how education and industry can ‘prepare for [AI’s] implementation in the production pipeline’ (The Animation Workshop/VIA University College, 2022: 12). Experts suggest that integration should begin with fundamentals: students need a solid understanding of core AI concepts and develop comfort with AI systems (Saam, 2022). Making AI literacy ‘second nature’ enables students to experiment with novel storytelling techniques that were previously inaccessible (Saam, 2022).
New research questions about AI and image processing in virtual production are emerging in academic literature (Swords and Willment, 2024: 12). A UK case study highlights that with emerging digital technologies like AI, workers must maintain relevant skills amid changing workflows (Willment et al., 2024a: 15). That is, Film schools should prepare graduates to master current VP techniques while adapting to future AI-driven tools.
Challenges in incorporating emerging technologies
Significant challenges moderate the current integration of AI and ML in virtual production education. What is taught in 1 year risks obsolescence in the next (Cote, 2024). VP’s breadth requires students to learn filmmaking, real-time 3D, and data science programming, areas where film faculties may lack expertise (Grossmann, 2024).
Notable skill gaps indicate incomplete integration. A 2023 industry report identified ‘substantial gaps in both technical and soft skills’ when technologies like programmatic scene creation and ML models enter production workflows (Digital Catapult, 2023a: 25). Professionals lack training in using AI/ML in virtual production, suggesting partial incorporation of these competencies in educational programs. The report identifies ‘develop or apply data mining and ML algorithms’ in real-time production as a requirement that few curricula cover (Digital Catapult, 2023a: 30). While future filmmakers must ‘use […] artificial intelligence tools’, this remains a future capability rather than a standard learning outcome (Digital Catapult, 2023a: 49).
Resource constraints and curriculum updates strain many film programs. Not all schools have access to AI software or machine-learning capabilities. Faculty need training in AI techniques and instructors with cross-disciplinary expertise. Therefore, programs should balance technical training with creative film education, as the UK case study suggested, by emphasizing ‘professional (transferable) skills’ alongside technical knowledge (Willment et al., 2024a). Educators should also integrate AI to enhance storytelling, because AI’s implications extend to psychological, cultural, and sociological impacts on society (Kavakli and Cremona, 2022).
Discussion
Uneven global access to virtual production education in English
The mapping shows that most identified VP programmes taught in English are clustered in the USA and UK, with Australia, Canada, and Ireland accounting for the majority. Continental Europe appears as smaller hubs, Asia has only a couple of programmes, and no English-language entries exist from Africa, Latin America, or the Middle East. The study also notes that using English as a search criterion shapes this pattern and that students in underrepresented regions may have to rely more on online training, partnerships, or international mobility to access training in VP workflows.
This pattern suggests that formal VP education in English is easier to find in well-resourced Anglophone systems and less visible to the broader community, or maybe less developed with an international profile elsewhere. While part of this reflects the method (focussing on English-language curricula), it also points to differences in who has the infrastructure, staff, and branding capacity to frame VP as a distinctive educational offer. Students in regions without such programmes are more likely to encounter VP through indirect routes rather than as a local, institutionally embedded pathway.
From a Bourdieusian perspective, 1 VP stages and associated infrastructures can be seen as a form of technological and symbolic capital. Institutions that can invest in LED volumes, real-time pipelines, and English-language marketing use these resources to position themselves as innovative and industry-facing within the wider field of higher education. Rather than a neutral technology, VP becomes one element through which universities differentiate themselves and signal prestige, drawing on existing reserves of economic and cultural capital.
This reading suggests that debates about VP implementation in HEI should pay attention not only to technical feasibility but also to how access to VP infrastructure may reinforce existing differences between institutions and regions. For practice and policy, these points toward the value of capacity-building strategies – such as shared facilities, targeted partnerships, or support for non-English and local-language programmes – that treat VP as a shared resource rather than a niche prestige asset. Future research could map informal and non-English VP training to see how these less visible spaces interact with, or offer alternatives to, the Anglophone university hubs identified here. Furthermore, future research could also examine curriculum transferability by exploring how VP programmes developed in well-resourced Anglophone institutions can be adapted to different linguistic, economic, and infrastructural contexts, and what changes are needed to make them work effectively in underrepresented regions or small film markets.
Engine-centric and technically weighted VP curricula
Across the mapped programmes, the tool and subject distributions show a clear engine-centred profile. Real-time and VP tools (such as Unreal, mocap, LED walls, XR platforms, and tracking) dominate the lists, with Unreal mentioned far more often than individual DCC tools. By contrast, audio tools and sound-related subjects appear only occasionally, and areas like scriptwriting, previs/storyboard, and game design are present but marginal. Overall, the strongest curricular emphasis falls on real-time engines and on-set technical operation rather than on story, sound, or game-oriented design.
Taken together, these patterns suggest that many VP curricula are organized around mastering the real-time engine pipeline and stage operation, with narrative development, sound, and game-adjacent design in supporting roles. The programmes appear to treat engines, tracking, and LED stages as the central spine of learning, with other domains added as integration points rather than coequal strands. This emphasis aligns with current industry demand for technically fluent crew, but it also risks narrowing students’ experience of VP to tool proficiency and on-set execution, leaving less structured space for story craft, audio, or experimentation with alternative workflows.
From a Latourian perspective, 2 these curricula can be read as efforts to align students with specific human–machine networks rather than with ‘technology’ in the abstract sense. VP pipelines – engines, LED walls, mocap systems, tracking rigs, DCC tools, and the people who operate and coordinate them – form actor-networks that structure what counts as competent action on a VP stage. When courses place real-time engines and stage operation at the center, they effectively configure students as particular kinds of actors in this network: reliable engine operators and pipeline nodes, more than authors of stories or designers of sonic space. Script, sound, and game design become weaker actors in the network, with less curricular weight to shape practice.
Reading VP curricula as human–machine networks highlights that educational design is already making choices about which parts of the pipeline are allowed to ‘speak’ most loudly in teaching. For practice, this suggests that programmes wanting more balanced graduates might deliberately strengthen underweighted actants – such as audio workflows, script and previs, or interactive/game design – so that the VP network students inhabit includes robust narrative and sonic decision-making, not only engine fluency. For research, this Latourian angle opens questions about how different configurations of tools, spaces (stages, labs), and assessment practices stabilize particular visions of VP work, and how alternative configurations might support broader creative and critical capacities around the same technologies.
Pedagogical convergence on ‘learning by doing real jobs’
Across programmes, a shared pedagogical pattern emerges. Undergraduate VP degrees use project-based learning and studio work, with storytelling and reflection in assessment. Postgraduate programmes emphasize specialization through practice-led, research-driven models, tied to interdisciplinary projects and industry internships. Certificates and diplomas compress VP training into short, employability-focused formats using work-based learning, studio partnerships, and portfolio assessments. Across all levels, a common logic emerges: authentic, production-like projects are central, under labels such as project-based learning, work-based learning, problem-based learning, and on-the-job training. Critical reflection, stronger at the postgraduate level, appears in vocational contexts, and block teaching, workshops, and demo-reel assessments mirror production cycles rather than traditional semesters.
These patterns suggest a strong convergence toward a ‘learning by doing real jobs’ model, in which teaching is organized around simulated or actual production work. Even highly vocational, employability-driven courses incorporate reflective and written components, indicating a degree of ‘academisation’ rather than pure skills training. At the same time, structuring time as blocks, sprints, and deliverables aligned with production cycles blurs the line between learning spaces and work-like environments. Students are positioned less as classroom learners and more as junior crew, moving through tightly scheduled tasks that resemble professional pipelines.
From Lefebvre’s perspective, 3 VP stages and associated teaching arrangements can be read as socially produced educational spaces rather than neutral containers. When curricula are organized around production-style blocks, demo-reel deadlines, and on-set studio practice, the VP stage is materially reorganized as a hybrid space where learning, labour, and institutional priorities are woven into real-time pipelines. The structuring of space (LED stages, labs) and time (intensive workshops, shoot-like schedules) reflects industry rhythms and institutional goals as much as pedagogical intentions. In this sense, ‘learning by doing real jobs’ is not only a method but a spatial reconfiguration that turns VP education into a miniaturized production environment.
Seeing VP pedagogy through Lefebvre reveals opportunities and risks. Designing educational spaces that mirror professional workflows can support transitions into industry and give students meaningful practice. However, if production logics dominate, critical inquiry, experimentation, and reflection risk being squeezed out by delivery schedules. For curriculum designers, this suggests the need to create spaces within VP stages not organized solely around deliverables, so students can question tools, workflows, and labour conditions while learning to operate within them. Future work could examine how VP programmes balance these dimensions and how students experience the stage as both a work environment and a learning space.
Competency stacks, career pathways, and structural gaps
The programmes describe VP learning in three broad bands: a shared technical core (real-time pipelines, immersive tech, 3D, mocap), creative competencies (story and content direction), and professional skills (collaboration, project management, emerging topics). Specializations extend these into areas such as engineering, creative filmmaking, virtual art department and stage operations, games/interactive media, and research or pedagogical roles. Audio and explicit leadership or supervisory skills are mentioned, but less frequently than technical and general creative outcomes.
VP education provides a clear structure: a strong shared technical base, space for story and creative direction, and an uneven layer of professional and leadership skills. The list of pipeline roles and engine-specific titles shows programmes align with a specialized VP labour market. The smaller presence of audio and lighter emphasis on storytelling and managerial skills indicate these competencies are less central to programme organization. Students are prepared to join the pipeline as operators or artists, but paths to audio-rich practice or supervisory roles are less mapped in curricula.
Following Lefebvre, these competency stacks can be seen as part of the social production of VP stages as educational spaces. The way learning outcomes, roles, and specializations are arranged effectively organizes the VP stage into zones of activity: areas where technical pipeline work dominates, places where story and content decisions are made, and smaller pockets where audio or leadership practices are developed. Rather than a neutral studio, the VP space is quietly shaped around the priorities of real-time pipelines, with story, audio, and coordination woven in but not always given equal spatial or curricular prominence.
This perspective underlines that designing VP curricula also means deciding which competencies are most visible at this stage. For practice, it supports approaches that build a shared core and then add clear specialization tracks, while giving story, audio, and leadership identifiable places in both the timetable and the studio workflow. Future work could explore how different spatial and curricular arrangements make certain roles – such as engine operator, director, producer, or sound designer – more or less imaginable to students as long-term career paths.
Early-stage integration of AI/ML and future-proofing VP education
Only a small number of programmes explicitly mention AI or ML in their VP curricula, for example, in relation to AI-driven animation, generative tools in coursework, or AI-supported avatars and digital doubles. Most course descriptions focus on engines, VP workflows and real-time rendering, with AI/ML present more in the wider literature on VP than in formal learning outcomes.
Together, this suggests that AI/ML is already important in VP practice but is only partially visible in current curricula. A few programmes are beginning to embed AI tools and workflows in a structured way, while many others still treat AI more as an emerging background technology than an explicit learning outcome. Educators are aware that tools change quickly and that AI will shape future pipelines. Still, they will face practical constraints around staff expertise, time, and infrastructure, and must also protect core film skills and creative focus.
From a Latourian perspective, AI and ML can be seen as new actants in the VP actor-network, joining engines, cameras, LED walls, motion-capture rigs, staff, and students. Where curricula explicitly integrate AI tools, they are beginning to reconfigure this network, allowing AI systems to take on visible roles in concepting, animation, and performance capture. Where AI is less explicit, the existing network remains centred on engines and stages, with AI present more implicitly through software updates and industry expectations. Curriculum design thus involves deciding how strongly AI/ML should be enrolled in the educational network and what kinds of human–machine relationships students will practise.
Viewing AI/ML integration in this way suggests that ‘future-proofing’ VP education is less about predicting particular tools and more about keeping the actor-network flexible. For practice, this points towards combining a stable VP core (engines, cinematography, story, collaboration) with growing opportunities for students to engage directly with AI-assisted workflows, supported by staff development and manageable resource investments. Emphasizing fundamentals, AI literacy, and learning-to-learn can help students adapt as new AI actants enter VP pipelines. Future research could track how different programmes bring AI into their stages – whether as occasional tools, embedded collaborators, or central organizing elements in teaching – and how this shapes students’ sense of working with intelligent systems in production.
Conclusions
The review argues that virtual production education is not just a neutral technical upgrade in higher education but a socially uneven, tool-weighted, and space-shaping formation that can reproduce institutional and regional advantages unless curricula are deliberately redesigned.
From this analysis, four key findings emerge. First, English-language VP programs are clustered in well-resourced Anglophone countries, with limited visibility elsewhere, partly due to the English-language search method and to unequal infrastructure and branding capacity. Even so, this fact highlights the difficulty of access for international students to English-trained curricula outside this realm and underscores the need for partnerships between HEIs to facilitate cross-cultural talent exchange. Second, the evidence shows a curriculum that is strongly engine-centric, prioritizing real-time pipelines and stage operations over story, audio, and game-adjacent design. Although aligned with industry practices and labour demand, this approach could imply a narrow conception of VP practice, without considering a broader spectrum of affordances across VP and real-time technologies. Third, pedagogy converges on ‘learning by doing real jobs’, reorganizing time and space around production-like blocks and deliverables, which strengthens industry transitions but may compress experimentation and critical reflection. Fourth, competency stacks map robust technical cores and specialized pipeline roles, while leaving structural gaps in story, audio, and leadership, with little evidence of AI/ML tools and formal outcomes despite the industry’s growing relevance.
Future research avenues.
Under the recognized evidence, beyond the standard quality assessment concerns about learning outcomes, and aligned with and in the same line of thought as the chosen theoretical lenses, this study claims that Virtual Production education must be evaluated as a field of institutional prestige (Bourdieu), a continuously evolving symbiotic human–machine network (Latour), and a socially produced learning space (Lefebvre).
This reframing shifts attention from whether programmes are just ‘well designed’ to how they distribute technological and symbolic capital, privilege certain competencies, and shape students’ identities as particular kinds of active pipeline actors.
Footnotes
Acknowledgements
The author acknowledges the use of ChatGPT-4.5 (OpenAI) for idea generation and exploration, language refinement, and coding assistance. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. This study was supervised by Dr Henrique São Mamede, Dr Arnaldo Santos, and Dr Johannes Steurer, to whom the corresponding author is highly grateful.
Ethical considerations
Participants provided verbal informed consent to participate, with approval/waiver granted by the Ethics Committee. The scenarios involving human participants were conducted within the scope of a PhD thesis, Web Science and Technology, reviewed and approved by the following institutions: the Open University and Trás-os-Montes and Alto Douro University.
Author contributions
R.A.S.: conceptualization, data curation, methodology, formal analysis, investigation, visualization, and writing:original draft preparation and resources, re-writing, reviewing, and editing. J.S.: technical supervision. H.S.M., & A.S.: funding acquisition and theorical supervision. All authors have read and agreed to the published version of this manuscript.
Funding
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
All data supporting the findings of this study are included in the article and its supplementary information. A complete replication package, including extracted study data and curricula datasets, is openly available in 
. Further materials are available from the corresponding author upon reasonable request.
Notes
Appendix
All data and resources used in the scope of this study were compiled and made available for replication. The packages in question are available online here:
Replication package:
Curricula data extraction: