Data Analytics and Artificial Intelligence in the Complex Environment of Megaprojects: Implications for Practitioners and Project Organizing Theory

Introduction

Project management can be summarized as the organized application of knowledge, competencies, and resources toward delivering project outcomes according to the acceptance criteria agreed in project parameters (Radujković & Sjekavica, 2017). Effective project management would result in the successful accomplishment of project deliverables while meeting the project scope requirements, financial obligations, time constraints, and stakeholder expectations (Radujković & Sjekavica, 2017). On the contrary, numerous negative impacts, such as monetary losses, the need for rework due to subpar quality, and damages to the organization's reputation, may result from inadequate or inefficient project management. Among different types of projects, extensive and complex multibillion-dollar endeavors executed over a longer time frame that affect the lives of millions of people are termed megaprojects and they are markedly different from regular projects in terms of complexity, characteristics, ambition, and politics (Pollack et al., 2018). Some examples of contemporary megaprojects include events, such as the Olympic Games, spacecraft projects such as the International Space Station, and infrastructure projects, such as the high-speed rail project of China (Flyvbjerg, 2014), all of which cost well over US$1 billion, agreeing with the general definition of a megaproject (Brookes & Locatelli, 2015; Denicol et al., 2021). Governments have become increasingly keen on undertaking megaprojects during their tenures since they are often considered showcases of prestige, culture, economic development, and the strength of a government (Van Marrewijk, 2017). Moreover, the successful completion of megaprojects brings the added advantages of consumer benefits via improved quality, increased usage of locally available resources against imports, and sustained employment opportunities (Flyvbjerg, 2014).

However, numerous complications and challenges associated with the successful execution of megaprojects have been identified (Eweje et al., 2012). The inherent difficulties in managing the Iron Triangle of a project—scope, time, and budget—can be found in megaprojects. The evaluation of megaprojects from this point of view would result in the majority of them being classified as unsuccessful (Erol et al., 2018). Cost overruns are a significant challenge that many prominent megaprojects, such as the 1976 Montreal Summer Olympic Games, the construction of the Sydney Opera House, and the Panama Canal, could not overcome (Flyvbjerg, 2014). Furthermore, schedule overruns or delays may also hinder the completion of megaprojects within the allocated time frame (Chen, 2019). These are further affirmed by Callegari et al. (2018), who conducted a study on Brazilian megaprojects in the power generation sector and revealed that nearly 76% of them encountered cost overruns, and 56% were completed with delays. Stakeholder management is another complication linked to megaprojects (Winch, 2017). For instance, concerns from the general public, especially indigenous communities, may arise because of the significant environmental footprint of these large-scale projects (Maher, 2019). The importance of stakeholder management has been highlighted by Lopez del Puerto and Shane (2014), who have stated cognition of the cultural and sociopolitical environment of the project and public outreach to be two success factors related to megaprojects.

Additionally, these large-scale ventures contain a higher level of risk despite their lucrative nature, arising due to the complexity associated with their immense scale, prolonged duration, sizeable finances, and the subsequent extensive planning required (Zidane et al., 2013). Hence, risk management is another integral aspect of megaprojects that should be given close attention. It is a process in which risks are identified qualitatively or quantitatively, and appropriate responses and contingencies are established along with risk monitoring and control mechanisms (A Guide to the Project Management Body of Knowledge [PMBOK^® Guide] – Sixth Edition, Project Management Institute [PMI], 2017). The most abundant risks studied in megaprojects are the schedule and cost overruns previously described, as asserted in the investigation by Irimia-Diéguez et al. (2014). Krane et al. (2010) identified 1,313 risk elements in seven megaprojects and found that 90% are operational risks directly related to the project's outcome. Short-term strategic risks that affect project features not explicitly mentioned in the project scope and risks impacting the clients and society are also essential to be considered as concluded in their study.

Megaprojects naturally involve more uncertainty and complexity due to their larger nature and the higher numbers of stakeholders and communication networks. Due to this, they are more likely to fail when compared to smaller projects (Flyvbjerg, 2014; Irimia-Diéguez et al., 2014). Hence, sound project management practices and sophisticated management skills are necessary to manage the ever-changing project environment, substantial logistical and supply chain aspects of the different component flows, and human resources (San Cristóbal et al., 2019; Srinivasan et al., 2021). Nevertheless, care should also be taken to avoid oversimplification (Giezen, 2012).

The absence of integrated teams may be problematic and even lead to the project's failure. As Edward Merrow has stated, “Developing an integrated team, one in which no key functions are in absentia, is the single most important project management practice for a megaproject” (Merrow, 2015c, p. 168). The author has also mentioned that the percentage of teams integrated into the front-end development phase is considerably low. In addition to integrating the teams, all relevant stakeholders in different project phases—initiation, planning, design, delivery, and closure—should also be integrated into the project. This calls for the requirement of governance (Mbassegue et al., 2017). Open governance of megaprojects would allow the responsibility of forecasting to be distributed, resulting in more realistic predictions (Mbassegue et al., 2017). It would also bring other advantages such as improved alliances between stakeholders, shared notions, and strong and widened knowledge bases.

Although the successful completion is of paramount importance, most of the recommended procedures are being ignored and bypassed due to their time and resource intensity (Merrow, 2015d). This can lead to errors and oversight, such as stakeholders’ misalignments, absence of realistic goals, and overly optimistic or pessimistic budgets and schedules in the early stages, which are the most common causes of project failures (Merrow, 2015b). Another critical aspect to be considered in the primitive stages is the availability and accuracy of primary data. Insufficient or inaccurate data is a significant factor in cost or schedule overruns, ultimately rendering projects unsuccessful (Love & Ahiaga-Dagbui, 2018; Merrow, 2015a). Hence, mechanisms and systems should be established to ensure the availability and reliability of data and information in the early phases.

During the megaproject's execution, cost overruns and delays are the most common hindrances expected, as described previously. Furthermore, the satisfactory delivery of the anticipated benefits has proven challenging. It has been discovered that only 10% of megaprojects are completed within the preagreed schedule, and it is the same for being on budget and delivering the intended outcomes (Flyvbjerg, 2011, 2014). When these three statistics are considered together, this will result in only 0.1% of megaprojects being deemed successful in terms of cost, schedule, and scope. This trend of predictable failure is known as “The iron law of megaprojects” (Walsh & Walker, 2021). Denicol et al. (2020) performed a comprehensive analysis of recent literature to identify reasons and solutions for megaprojects’ underperformance. The authors suggest the involvement of the manufacturing industry, communities, and institutions within megaprojects and enhancing and integrating system architecture as solutions. Carrying out project work in an orderly way and establishing comprehensive yet pragmatic work schedules can help avoid delays (Merrow, 2015e). It is also recommended to prevent fixed-price contracts due to the substantial probability of cost overruns (Parth, 2014). Moreover, in an overrun, it is advised to avoid labeling the superficial differences between the planned and the actual values and assess them thoroughly in a broader context (Walsh & Walker, 2021).

Data analysis is gradually becoming an indispensable aspect of megaproject management due to several reasons (Zangeneh & McCabe, 2020). Most megaprojects would essentially be the first time a particular venture would be implemented in a specific geographic area impacting its environment, economy, and people; hence, there would be negligible or no prior experience for project personnel (Litsiou et al., 2019). This essentially leaves them unable to depend on past data and calls for structured and judgmental forecasting techniques to predict outcomes based on various information and scenarios. Moreover, the emergence of information technology has radically enhanced the quantity of data and information available and the number of different methods of analyzing large swaths of data efficiently and effectively. Hence, incorporating modern data analytic methods has become crucial for improving megaproject performance (Sarkheyli & Sarkheyli, 2019).

The use of data analytics can be especially seen in risk assessment, where analytic techniques, such as Monte Carlo simulations, the three-scenarios approach, and sensitivity analysis have been effectively used for quantitative risk analysis (Nabawy & Khodeir, 2020). For instance, Purnus and Bodea (2013) applied the Monte Carlo method in a construction project scenario. They concluded it to be far more accurate than the three-scenarios approach despite its relatively high complexity and time-consuming nature. Other than risk management, these practices can also be used for forecasting likely future outcomes through the analysis of historical data and identifying trends. Earned value management is the most commonly used technique for forecasting project outcomes (Batselier & Vanhoucke, 2015). It uses the current cost and schedule status to accurately predict future results, enabling managers to initiate necessary control measures. Earned duration management and earned schedule are methods derived from earned value management. However, it is noteworthy that earned value management cannot analyze information from past projects to understand the outcome of the considered project before its commencement; it is only used in the middle of the project to assess the progress and make predictions (Litsiou et al., 2019).

Currently, the world is undergoing the fourth Industrial Revolution—colloquially known as Industry 4.0 or 4IR—which emphasizes the interconnectivity of different systems, automation of construction and production environments, real-time information, and machine learning and artificial intelligence (AI) (Maskuriy et al., 2019). Digitization of the information ecosystem is also transforming project delivery. Many projects are rapidly adapting information digitization due to higher levels of integration, interconnection, innovativeness and the easier dissemination of digitized knowledge and information (Boland et al., 2007; Kache & Seuring, 2017). These data can be accessed remotely and searched rapidly. Problems can be identified quickly because the information is available in real time (Whyte, 2019). These inherent benefits have been proven to be vital, especially during the ongoing COVID-19 pandemic, for studying the movement of people and predicting and modeling outbreaks (Gadhi, 2020). With the adoption of this digital delivery in the context of megaprojects, numerous types of three-dimensional (3D) models, geographic information systems, and databases are created (Whyte et al., 2016). These available databases present an opportunity to identify reasons for the gaps between expectation and actual project performance using data analytics and further improve project management (Bardhan et al., 2007; Harty & Whyte, 2010).

There is still considerable work regarding the widespread use of data analytic methods in a megaproject context. Hence, the objectives of this review article are (1) to critically evaluate the current progress of data analytics into megaprojects, (2) to identify how data analytic techniques can be further developed into more advanced tools to be used in construction management, and (3) to suggest recommendations for future megaproject management practices.

Review Methodology

This article presents a detailed review of available digitalization tools and data analytic methods, their applications in contemporary projects, and how they can be adapted to larger ventures such as megaprojects. Also, the benefits of utilizing these tools for megaprojects, potential challenges that may arise during this transformation, and likely future developments and research directions have been discussed.

A comprehensive literature survey was carried out to gather the literature in this field, and a keyword search was performed in scientific publishing databases such as “SAGE Journals,” “Google Scholar,” and “Compendex Engineering Village.” Specific keywords included—but were not limited to—combinations of pertinent terms such as “project management,” “megaprojects,” “fourth industrial revolution,” “digitization in megaprojects,” “data analytics,” “digital project delivery,” as well as of specific technologies such as “artificial intelligence,” “automation,” “information modeling and digital twins,” and “life cycle thinking in megaprojects.” Articles published after 2010 were selected from the literature survey results and prioritized to be reviewed to maintain the relevance of the study.

Artificial Intelligence in the Complex Environment of Megaprojects

Overview of Digital Delivery of Megaprojects

Digitalization of megaprojects has numerous inherent advantages, such as analyzing past information and foreseeing unfavorable circumstances, performing hypothetical what-if scenarios, and automating critical elements such as management of human resources, financial resources, and risks. An overview of the digitalization of megaprojects and construction projects is illustrated in Figure 1 (Sawhney et al., 2020). As evident, information modeling (IM) and cloud-based computing would be the cornerstones of the digitalization framework, where the former would facilitate operational research as well as modeling and simulation of the project environment (Sawhney et al., 2020), whereas the latter would provide secure and unlimited storage for data and information (Cooper, 2018). These two key technologies facilitate the integration of physical and virtual layers of megaprojects throughout their duration. Moreover, there have been multiple means of utilizing digital technologies and facilities in projects, as discussed in subsequent sections.

OPEN IN VIEWER

Figure 1.

Overview of digitalization of megaprojects (Sawhney et al., 2020).

Digitalized technologies analyze the data and information obtained from previously executed projects or ongoing project in a digital environment, facilitating accelerated and accurate decision-making for key stakeholders (Oesterreich & Teuteberg, 2016). During the preliminary stages, digital tools are most commonly employed for planning purposes via extracting and analyzing information from previous ventures of a similar nature. However, as projects get underway and gradually move forward, the focus of digitalization would shift toward monitoring and control. This pensive digitalization would benefit megaprojects in several ways such as increased sustainability via efficient resource consumption and waste reduction, cost reduction, on-time completion, increased on-site safety, enhanced outcome quality, and reduced uncertainties (Oesterreich & Teuteberg, 2016).

Recent Progress in the Digitalization of Megaprojects and Construction Projects

Table 1 summarizes recent studies involving digitalized technologies in megaprojects and other construction ventures. Risk assessment has been one of the core objectives of incorporating these technologies into projects. A wide range of risk types has been considered in these studies, including information risks (Kovtun et al., 2020; Kuznetsov et al., 2021), and occupational safety incidents (Ayhan & Tokdemir, 2019a, 2019b). Other objectives focused on in academic studies and projects include energy management (Cho et al., 2015), enhancing efficiency (Aziz et al., 2014), stakeholder management (Gaur & Tawalare, 2022), and analyzing public opinion (Vinichenko et al., 2021; Zhou et al., 2021). Optimization techniques, such as quasi-Newton algorithm and genetic algorithm, are tools employed in various projects for optimized and efficient allocation of resources and time (Kaiafa & Chassiakos, 2015). They have also played a vital role in recent studies in optimizing construction schedules, physical and financial resource usage, and maintaining the quality of construction projects (Aziz et al., 2014; Faghihi et al., 2014).

OPEN IN VIEWER

Table 1.

Recent Studies Focus on Using Digitalization Technologies in Megaprojects

Reference	Technique/Description	Novelty/Outcomes/Key Findings
(Zangeneh & McCabe, 2020)	Creation of data infrastructure for analytics of megaprojects using semantic web	Unified ontology for knowledge representation and risk assessment in megaprojects
(Aziz et al., 2014)	Critical pathway of scheduling integrated with Genetic Algorithm	Improved resource utilization Faster completion Financial optimization Improved quality
(Faghihi et al., 2014)	Using geometric information from BIM in Genetic Algorithm for generating and optimizing scheduling	Use of geometric information from BIM toward optimized construction schedules
(Gaur & Tawalare, 2022)	Studying the involvement of BIM in stakeholder management in a metro-railway megaproject	Important stakeholder management elements, such as communication, integration, risk management, and knowledge sharing can be fulfilled via the benefits of BIM.
(Zhou et al., 2021)	Use of Latent Dirichlet allocation for topic modeling and sentiment analysis in megaprojects	Analysis of online public opinion via social media information can be conducted for megaprojects.
(Vinichenko et al., 2021)	Assessing the inclination of megapolis residents to partake in megaprojects involving the digital economy and AI	The willingness depends on numerous factors such as age, nature, and attractiveness of the megaproject. Participants were motivated by prospects of salary, networking, and experience.
(Kifokeris & Xenidis, 2019)	Development of a framework to evaluate the constructability of a project with known risks using supervised machine learning	Creation of an accurate classifying equation for forecasting the constructability of a project integrating risk analysis
(Kovtun et al., 2020)	Consideration of information risks for developing loyalty programs for megaprojects	Convolutional neural networks were capable of risk analysis with higher accuracy, thus more suitable for loyalty programs,
(Lobo & Abid, 2020)	Analysis of impact on public and resistance to a railway megaproject via social media	Recommended to utilize social media marketing principles for enhancing relationships among stakeholders
(Natarajan, 2022)	Use of machine learning for predicting cost and schedule overruns in oil and gas megaprojects	Reference class forecasting results were improved through the machine learning model via learning the relationship between project features and outcomes.
(Kuznetsov et al., 2021)	Assessing information risks using neural networks in semantic information fields in megaproject life cycles	Information risk management and development of loyalty programs is possible via semantic information fields.
(Park et al., 2017)	Use of BIM, bluetooth low-energy-based location, and cloud communication for safety monitoring	Enhanced construction safety via a novel automated monitoring system with low cost
(Cho et al., 2015)	Cloud-based software for energy management decision-making of construction projects	Anticipated optimization of energy consumption in construction projects
(Ayhan & Tokdemir, 2019a)	Prediction model for safety incidents in construction using artificial neural networks and addressing uncertainty with fuzzy sets	84% of outcomes forecasted with 90% confidence
(Ayhan & Tokdemir, 2019b)	Prediction model for safety incidents in construction using artificial neural networks and case-based reasoning, and using latent class clustering analysis for maintaining homogeneity	Case-based reasoning performed better than artificial neural networks

Challenges of Implementing Digital Tools in Megaprojects

Regardless of the potential for innovation and the convenience the digitalization can bring to megaprojects, there would undoubtedly be numerous challenges and setbacks during their implementation stage and once they commission. Figure 2 illustrates several challenges that could arise while establishing these technologies in megaprojects.

OPEN IN VIEWER

Figure 2.

Challenges in digitalization of megaprojects.

One key obstacle would be skepticism and resistance to change, which would be due to the conservative viewpoints of senior project leaders (Majrouhi Sardroud, 2012; Sawhney et al., 2020). Several other factors might fuel this resistance, and one such factor would be the high initial investment needed. This increased capital investment would comprise equipment and technical components, as well as costs of providing necessary training for project personnel, legal expenses, and consultation fees in the primary phases (Akanmu & Anumba, 2015; Smith, 2014). Despite training the existing labor force, acquisition of specialized and skilled personnel may be required during the transition to digitalization and might even create resistance within management. In addition, modernization of project environments via the adoption of novel technological tools would call for changing and restructuring existing practices, which may be seen as a hindrance to existing operations and thus opposed (Oesterreich & Teuteberg, 2016; Owen et al., 2010).

Furthermore, most of these digitalization tools are currently in their preliminary development stages in the context of megaprojects. Consequently, the absence of proper standards, reference architecture, and regulatory and legal precedents impedes their adaptation (Lasi et al., 2014; Oesterreich & Teuteberg, 2016). Additionally, the implementation of these technologies would mean the necessity of tracking the movements of project personnel for analysis purposes, which raises several ethical concerns on privacy and personal space (Smith, 2014). Obtaining consent from tracked individuals, giving them control to some extent in having their movements tracked, and discussing when and to whom this location information would be disclosed would be necessary before implementing these technologies (Liao et al., 2012).

Cybersecurity has become a vital element of any digital platform (Untawale, 2021). Hackers may try to obtain critical information via cyberattacks, which they would make public, attempt to sell to other organizations, or demand ransoms from the attacked entity itself (Troia, 2020). Megaprojects, like construction projects, may be specifically vulnerable to such attacks due to the large swaths of information in digital databases such as workers’ personal information, design blueprints, and financial records (Parn & García de Soto, 2020). As previously discussed, this would also be linked to the ethical principles of tracking worker movements. Tracked location information should be secure enough not to be accessible by malicious factions (Liao et al., 2012). Hence, the security of data fed into and stored in the megaproject data network is imperative.

The Possible Future of AI in Megaprojects

The cyberphysical systems and digitalization technologies discussed have been integral elements of the fourth Industrial Revolution (Afolalu et al., 2021). Examples include ride-hailing and sharing apps, such as Uber and Lyft, and online retailers like Amazon. They are credited with drastically revolutionizing customers’ busy lifestyles in various ways (Kirkpatrick, 2018; Shankar, 2021). There is enough evidence that digitalization has become inexorable. Not managing to stay up to date with the latest concepts and technologies may result in even leading organizations being driven into obsolescence, as in the cases of Nokia mobile phones and Kodak cameras (Hodges & Gill, 2015; Vuori & Huy, 2016).

The digitalization of megaprojects is also inevitable, and digital technologies have been steadily extending into project management in many industries (Aghimien et al., 2020). Advantages of digitalization of construction projects include reduced risks and wastage, improved allocation and sharing of resources, and enhanced performance (Aghimien et al., 2020), improving overall project delivery. Furthermore, this would be beneficial for supply chain management in terms of bridging geographical and cognitive distances—colloquially referred to as the proximity—between different project parties, resulting in increased harmonization of suppliers with the project (Dallasega et al., 2018). However, the progress of digitalization in this area is one of the lowest thus far (Gandhi et al., 2016); hence, amplification of research and pilot projects in this area is desirable.

 Figure 3 illustrates the immediate steps forward in the digitalization of megaprojects. Integrated project delivery (IPD) is a conspicuous emerging trend seen in the construction industry. It enables a higher level of partnership among the organization, stakeholders, systems, and practices to utilize the core competencies of each party more effectively, ultimately improving the efficiency of project work while eliminating waste of resources (Lahdenperä, 2012). A key feature of IPD is the early participation of all key stakeholders from the stages of planning, defining goals, and allocating resources (Jones, 2014), resulting in shared benefits for all parties. Moreover, it has been proven that IPD can curb the cost of project alterations by 33% (Trach et al., 2019). In addition, there is immense potential for using IM and BIM to support this integration among involved parties at different stages (Koc et al., 2020). Thus, adapting IPD models supplemented with digitalization for megaprojects should be an important future research direction.

OPEN IN VIEWER

Figure 3.

Prospects of digitalization in megaprojects.

Koc et al. (2020) have presented the evolution of numerous digitalization technologies in the field of construction over time, defining five status quo levels beginning with awareness (educated only about the general idea, not specific details) being the least advanced and adoption (large-scale implementation) being the most. According to their assessment, only IM, the Internet of Things (IoT), and cloud computing have progressed sufficiently to be close to the adoption stage between the 1960s and 2019. Nonetheless, it has been foreseen that accelerated work digitalized project delivery along with integration will enable almost all technologies to be progressed to trial and adoption stages before 2040, with IM, IoT, cloud computing, mobile interfaces, big data analytics, and large-scale additive manufacturing being firmly established in the industry (Koc et al., 2020). Thus, it is fair to forecast that these well-established concepts and technological platforms will make their way into a much larger scale in megaproject ventures.

Education would have to play a paramount role in materializing these revolutionary changes. Engineering edification needs to consider emerging digitalization trends, transition AI capabilities, and reshape its programs accordingly. Moreover, these programs would need to have a considerable component of interdisciplinary study and research, since digitalization would require engineers to step beyond their core competencies into diversified avenues. Undoubtedly, this will be a crucial skill in the future (Knippers et al., 2016; AbouRizk et al., 2011). Thus, the on-site implementation of such technologies is essential to experience their perceived benefits firsthand.

Increasing the availability of data and information would be advantageous in supplementing an accelerated transition into digitalized megaprojects, especially in technologies involving data analysis (Koc et al., 2020). The availability of sufficient data and information is the basis of many digitalized technologies, including AI (Sundblad, 2018). Moreover, to enhance the completeness and thoroughness of such systems and bridge the physical and digital ecospheres, it is essential to incorporate the entire life cycle of such projects into them (AbouRizk et al., 2011). Integration of all project phases using IPD can be beneficial in adapting this life cycle thinking approach into megaprojects (Koc et al., 2020), ultimately enabling key stakeholders and decision makers to assess the life cycle impact of their plans and arrive at better conclusions.

AI technologies have improved over recent years, having been expanded substantially in terms of available learning algorithms, enhanced computational power, and increased availability of extensive datasets (Brynjolfsson et al., 2018). With the introduction of advanced digital technologies and automation in management, low-skilled labor performing routine project management tasks is highly likely to be replaced (Parn & García de Soto, 2020; Wilson & Daugherty, 2018). Nonetheless, the need for a highly skilled workforce in data analytics and AI will manifest who can understand customer or project AI needs, work on the synthesis of strategic project roadmaps for AI systems, and help foster better project delivery (Wilson & Daugherty, 2018). Furthermore, digitization measures and AI capabilities are essential as participants’ interests and the future of work evolve. Lastly, the security of the digital infrastructure and the AI system and attending to AI’s responsibilities need to be considered seriously and need further exploration and evolution (Parn & García de Soto, 2020; Walters & Novak, 2021). This is essential since cybersecurity compromises can be escalated to national security threats, as seen in recent incidents (Ventre, 2020; Walters & Novak, 2021).

Theorizing Data and AI in Megaprojects with Information Economics

Many theoretical frameworks have been applied to megaprojects, including behavioral economics (Flyvbjerg, 2021b; Flyvbjerg et al., 2009; Steen et al., 2017), organizational capability theory (Davies et al., 2016; Steen et al., 2021), and institutional theory (Biesenthal et al., 2018; Brunet, 2021). However, theorizing how data and AI will change the organization of these projects requires an understanding of the costs and value of information and the economics of organizing projects. We suggest that the field of information economics is a fertile area for interpreting the impact of AI. In this section, we give a brief overview of information economics and then use current discussions on the economics of AI to connect with the extant literature on the theory of the firm and project-based organization.

Information economics is a branch of economic theory that has origins in a 1945 seminal essay by Hayek (1945) on knowledge and the superiority of market economies over central planning. This article inspired several economists to consider the economic properties of information, most notably Stigler (1961), who noted the peculiar absence of the role of information in conventional economic theory. Following Stigler, the field of information economics was developed as a core part of modern economic theory, most notably with contributions from Nobel laureates Arrow (1996) and Stiglitz (1994).

The information economics literature now has several branches, but three of them are most pertinent to theorizing AI in megaprojects. These are the value of information (Feltham, 1968; Howard, 1966; Samis & Steen, 2020), information as an economic good (Arrow, 1974; Lamberton, 1997), and the network effects of information-based business (Shapiro & Varian, 1999) that causes increasing returns to scale. To summarize the application of information economics to the subject of AI in megaprojects, we will advance propositions that may inspire future research.

AI increases the value of information by enabling the prediction of outcomes from decision choices (Agrawal et al., 2018). In classical value of information theory, this is called clairvoyance (Howard, 1966). Information may always have some predictive value in megaprojects, but AI increases that value when combined with information derived from related circumstances. The value of this related information will not increase linearly as it is gathered. Instead, when it crosses the threshold of supporting AI, it grows exponentially. For megaprojects, modularization can increase the value of information by standardizing aspects of megaproject design in replicated modules and allowing the generation of similar information compared with one-off designs. In this way, reaching information thresholds for implementing AI will happen in more instances with modular design and construction. Modularization has been suggested to be a way to improve megaproject performance (Flyvbjerg, 2021a), but AI will accelerate this move to modularization, hence Proposition 1:

The increasing value of information-driven AI use in megaprojects requires theory that considers information as a valuable economic good, in the way that other inputs to project delivery are valuable such as people, materials, and technology. However, information economists understand that information is an unusual good in the sense that it is difficult to price with confidence until the buyer actually has it (Arrow, 1974; Lamberton, 1997). In other words, markets for information are imperfect because of information asymmetry. A megaproject owner might find it difficult to assess the relevance of AI-based on information from a different project managed by another organization. From a transactions–cost view, this will increase the economic efficiency of integration and quasi-integration strategies, such as alliances, allowing organizations to be design and delivery partners across multiple projects (Dietrich, 1994; Williamson, 1990). The increasing value of AI for information will also accelerate the move to implement data capture and storage standards to reduce the uncertainty around the importance of information. Based on these ideas, Proposition 2 is stated as follows:

There is a long tradition in the fields of strategic management and organizational economics that considers the impact of technology on organizational forms (Zammuto et al., 2007). With projects now clearly recognized as a form of organization, there are many opportunities to theorize the future of AI in projects through this lens.

Conclusion

In this article, we have reviewed the literature on the impacts of digitalization and AI on projects. These changes will have significant implications for managing megaprojects, and the prospect of improved delivery outcomes for megaprojects is substantial. The following final remarks on the future impacts of digital technology can be made.

Cost and schedule overruns are inherent issues in most megaprojects. Digitalization can help project managers interpret a large volume of data and suggest an efficient approach in real scenarios for resource utilization to avoid such overruns and resource depletion. Optimized digital models have the prospect of maintaining the duration and cost of megaprojects in check while maximizing the output quality. Risk management can also be executed by calculating the risk indices in real time throughout the project. The data analytics and predictive analysis can foresight the project’s outcomes either in a run or yet to be started. Digitalized tracking tools can aid in optimizing resource allocation to reduce costs while maintaining the schedule. The insights from the tools discussed can support effective project management.