Vol. 12, No. 08 [August 2026]

The construction sector accounts for roughly 13% of global gross domestic product and employs about 7% of the world’s workforce, yet it remains among the least digitised industries and continues to record chronic cost and schedule failure. Artificial intelligence (AI) is widely promoted as the corrective, and the volume of scholarship has grown steeply from fewer than ten journal articles annually before 2018 to 125 in 2025 within the cost, time and safety domains alone. What this has not resolved is the relationship between reported model accuracy and realised project performance. This paper critically reviews the peer-reviewed evidence on AI-enabled digital project management (DPM) in construction, asking where AI is being applied across project management knowledge areas, what performance outcomes are actually demonstrated, and what impedes organisational adoption. Thestudyfirstly reported predictive accuracy being high and, in cost estimation, approaching saturation, with coefficients of determination above 0.99 now routinely claimed. Second, that accuracy rests on fragile foundations validation are overwhelmingly retrospective, samples are small and geographically bounded, and where hold-out testing is disclosed, the degradation is severe, one delay-prediction study reporting 47.2% testing accuracy against 74.5% on training. Third, adoption barriers cluster in the organisational domain, and the research literature systematically under-attends them: bibliometric analysis of 4,668 publications found socio-technical barriers occupying 4.6% of the literature against 46% practitioner salience, a ratio of 4.7 practitioners per publication. Study argues that these are one phenomenon rather than three. Predictive accuracy on retrospective single-context data has become a poor proxy for managerial value, and the field’s dominant metric now obscures more than it reveals. An integrated framework is proposed in which technical capability is mediated by deployment maturity and organisational absorptive capacity, and in which feedback latency operates as an under-theorised determinant of diffusion.
Keywords: artificial intelligence; digital project management; project performance; machine learning; technology adoption

Within the framework of European Union law, as the functioning of the internal market, public administration, essential services and democratic institutions across the EU increasingly depends on digital networks and information systems. The Union regulation always operates in connection with national legal frameworks. The Union regulation always operates in connection with national legal frameworks. This is achieved not only through the mere transposition of directives, but also through Member States‘ active engagement in the creation of their own legislation that respects national specificities as well as political or security priorities, The goal of this paper is to clarify this model on the example of the Czech legislation as adapted to the requirements of the EU‘s framework for harmonization. We focus on the New Czech Cybersecurity Act, No. 264/2025 Coll, adopted in November 2025. As regards the legislative method employed, the Czech Act makes use of the so-called ‗performance-based‘ approach and introduces a number of vague terms into the legal rules, which places increased demands on both the regulator and the regulated entities.
Keywords: cybersecurity law; EU cybersecurity legislation; New Czech Cybersecurity Act; performative method