{"id":905109,"date":"2025-12-11T13:53:43","date_gmt":"2025-12-11T20:53:43","guid":{"rendered":"https:\/\/www.myconstructionexpert.com\/blog\/?p=905109"},"modified":"2025-12-11T13:53:46","modified_gmt":"2025-12-11T20:53:46","slug":"artificial-intelligence-in-construction-while-mitigating-risk","status":"publish","type":"post","link":"https:\/\/www.myconstructionexpert.com\/blog\/artificial-intelligence-in-construction-while-mitigating-risk\/","title":{"rendered":"Automation and Artificial Intelligence in Construction: How to Capitalize on Technological Advancements While Mitigating Your Risk"},"content":{"rendered":"\n

Vivien Peaden and Jordyne Johnson Richartz | Baker Donelson<\/a><\/p>\n\n\n\n

(1) The Good, the Bad, and the Ugly of Artificial Intelligence in Construction<\/h4>\n\n\n\n

The commercial construction industry is undergoing a seismic shift as artificial intelligence<\/a> (AI) tools rapidly reshape how projects are planned, managed, and delivered. Advancements in AI and machine learning, including those in generative AI (GAI), robotics, augmented reality (AR), and virtual reality (VR), are streamlining every stage of the engineering and construction processes, from preconstruction to project completion. AI is presently being used to prepare cost estimates, develop bids and project proposals, compare as-built and as-planned conditions, monitor schedule progress, perform inspections, identify design conflicts, and track material prices, among other things. The automation of these processes not only boosts project efficiency but also improves the sector’s post-COVID-19 productivity challenges.<\/p>\n\n\n\n

According to Autodesk, a leading provider of construction management software, a recent\u00a0survey<\/a>\u00a0shows that the “AI in construction” market is expected to grow from around $11.1 billion in 2025 to $24.3 billion by 2030. Despite the advantages that AI affords contractors, engineers, and similar professionals, it is imperative to understand and mitigate risk as the construction sector experiences a gradual but transformative movement towards digitalization and AI adoption.<\/p>\n\n\n\n

(2) AI Is a Tool, Not a Replacement for Professional Judgment<\/h4>\n\n\n\n

AI should be used by contractors, engineers, and design professionals to augment their roles in the planning, design, and building phases \u2013 not to supplant their own professional judgment and expertise. The American Society of Civil Engineers’ (ASCE) Code of Ethics instructs civil engineers to “consider the capabilities, limitations, and implications of current and emerging technologies when part of their work.”\u00a0ASCE Code of Ethics, Sec. 1(h)<\/a>. As the incipient legal landscape governing the use of AI in the construction industry develops over the coming years, the policy statements of professional associations and trade organizations serve as a present guidepost.<\/p>\n\n\n\n

In July 2024, ASCE released a policy statement emphasizing an engineer’s unwavering ethical duties and professional responsibility, notwithstanding the adoption of AI tools. ASCE cautions that “AI cannot serve as a replacement for the professional judgment of a licensed Professional Engineer” and that such technology is no substitute for an engineer’s training and experience. See\u00a0Policy Statement 573 \u2013 Artificial Intelligence and Engineering Responsibility | ASCE<\/a>. The message to civil engineers and the greater construction industry is clear: engineers will be held accountable for the data upon which they base their decisions, recommendations, and representations, regardless of whether such data was computer-generated. Id.<\/em>\u00a0An engineer \u2013 not the AI application used by the engineer \u2013 will ultimately be responsible for violating any applicable standard of care. Id.<\/em><\/p>\n\n\n\n

The National Society of Professional Engineers’ (NSPE) February 2025 policy statement provides that engineers using AI are held to the same professional licensure standards as traditional engineers for purposes of public safety, health, and welfare. See\u00a0NSPE Position Statement No. 03-1774<\/a>. Similarly, the National Council of Architectural Registration Boards (NCARB) has stated that\u00a0“[a]ny\u00a0proposed regulation that addresses AI usage in practice must ensure the licensed practitioner remains in responsible control and continues to be accountable for all technical submissions under their seal.” See\u00a0October 17, 2024, NCARB Press Release<\/a>. In other words, AI tools cannot stand in for an architect’s or professional engineer’s stamp on drawings.<\/p>\n\n\n\n

(3) Privacy and Data Governance Concerns With AI<\/h4>\n\n\n\n

AI has significantly changed how the construction sector approaches project planning and field operations. Integrating AI into commercial construction workflows and processes means that companies are handling unprecedented volumes of data that boost forecasting, site safety, and productivity. For example, AI-enabled computer vision technologies provide real-time site monitoring that alerts site managers whenever workers are not wearing protective gear or when equipment is malfunctioning in violation of standard protocols. Further, GAI also delivers quicker outputs, enabling faster conceptual model development and accelerating design standardization. The speed and scale of AI adoption require organizations to adopt robust AI governance concerning what data they have, how new technologies are integrated, and who is accountable for outcomes.<\/p>\n\n\n\n

(a) “Black Box” AI<\/strong><\/p>\n\n\n\n

 All commercially available large language models (LLM), including OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, and xAI’s Grok, share the common problem of the “black box” myth, meaning that users can see their inputs and the resultant outputs without clear insight into how the model arrived at its conclusions. This lack of transparency has led to hesitation by some companies to adopt LLMs more broadly in their business operation. According to recent research by Anthropic in March 2025, Claude and other LLMs are capable of “hallucinations,” or false results, in generating their chain-of-thought in order to please a user. This finding makes it paramount for organizations to implement technical and operational safeguards to verify an AI model’s outputs before relying on such information.<\/p>\n\n\n\n

These transparency and interpretability requirements are further reflected in Article 13 of the EU AI Act, which mandates that providers of high-risk AI systems disclose their limitations (such as accuracy, robustness, and cybersecurity) and intended purpose, in addition to providing appropriate measures for implementing human oversight. Where AI is used for safety purposes as prescribed under certain EU regulations, these risk management requirements concerning high-risk AI use are set to take effect in August 2026 \u2013 e.g., 24 months after the EU AI Act officially went into effect.<\/p>\n\n\n\n

(b) “Shadow AI”<\/strong><\/p>\n\n\n\n

Shadow AI is becoming a growing issue for organizations that are still experimenting with AI pilot programs. For decades, “shadow” adoption of emerging technologies has plagued enterprise security teams, from sales departments storing customer data in unauthorized Dropbox accounts to the creation of rogue Excel spreadsheets. Today, it has evolved into something far more powerful called “Shadow AI,” which refers to the unsanctioned use of AI applications by an employee to perform work-related tasks without formal oversight by an employer’s procurement, IT, or compliance review teams.<\/p>\n\n\n\n

According to an MIT survey, while only 40 percent of companies have officially adopted LLM subscriptions, more than 90 percent of workers report regular use of personal AI tools for work. While employees’ unauthorized use of public AI tools can drive productivity, such use can simultaneously expose an organization’s proprietary information (including pricing information and sensitive personal data), waive privilege, and result in breaches of contractual obligations.<\/p>\n\n\n\n

(c) Data Leaks in Using Open-Source AI<\/strong><\/p>\n\n\n\n

Shadow AI often begins subtly, introducing significant risks such as data leaks or exposure to malicious AI models. Some examples include the following:<\/p>\n\n\n\n