As a leader in infrastructure engineering, Egis stands out for its ability to anticipate and integrate technological advancements, especially in the field of Artificial Intelligence (AI). Let's consider the transformative influence that AI could have on the design, execution, and management of infrastructure projects. The essential question is to determine to what extent the traditional aspects of infrastructure engineering can resist or adapt to the emergence of AI-based technologies.
The path to digitalization and automation
The transition to digitalization, exemplified by the widespread adoption of Building Information Modeling (BIM), has already revolutionized Egis's work methods. Current forecasts suggest that AI could automate a significant proportion (up to half of the processes in infrastructure engineering). This prospect raises crucial questions about elements of creativity, teamwork in complex contexts, and adaptability to the specific and unpredictable challenges encountered in infrastructure projects.
With the advent of Large Language Models like GPT-4, we are witnessing revolutionary advances that redefine the possibilities in our sector. These models offer unprecedented opportunities to optimize and streamline design plans. However, they also highlight the invaluable worth of human collaboration, especially in managing unforeseen and complex situations on sites, where human presence and intuition remain fundamental.
Limits of originality and innovation in AI
AI, capable of emulating existing architectural or structural models, nevertheless faces limits in terms of originality and innovation. Original design remains deeply rooted in human creativity, driven by societal needs and a man-directed vision. Although AI is highly efficient when it comes to following defined objectives, true innovation in the field of infrastructure requires a human perspective, combining intuition and understanding of complex societal needs.
In the context of infrastructure projects, the role of generative AI can become significant, but it requires rigorous human supervision for its initiation, validation, and adjustment. Emphasizing here the importance of data quality: the principle "Quality in, quality out" is fundamental. For infrastructure projects, the quality and reliability of input data are crucial to ensure the efficiency and accuracy of the AI-generated results.
The human at the heart of engineering roles
The advances in AI, certainly impressive, will constantly require human intervention and supervision, particularly in unpredictable situations on the field. Human expertise, with its ability to adapt and respond to unexpected challenges, remains an irreplaceable component in the field of professions.
Infrastructure projects, with their unique and often unprecedented challenges, demonstrate that AI cannot always generalize or apply uniform solutions. This limitation underscores the importance of human understanding of specific conditions and contexts.
Anticipating a potential slowdown in Moore's Law by 2025, we could face limitations in AI's ability to manage increasingly complex infrastructure projects. We would reach a technological plateau.
The vast amount of data required for AI training raises major concerns in terms of sustainability and environmental impact, given the energy-intensive nature of these processes. We must integrate this into our eco-design models.
Despite the increasing possibilities for automation or assistance in certain aspects of infrastructure engineering, complex challenges and current technological limitations will likely ensure the centrality and importance of human intervention in this field for at least another decade. The synergistic collaboration between human intelligence and generative AI opens pathways for innovation and efficiency improvement in infrastructure engineering, while respecting the unique constraints and needs of each project.
