Bricks and bots: AI technologies’ growing impact on construction

This article was originally published in the Daily Journal of Commerce on April 17, 2025.

Seemingly everywhere you look, there are stories about how the increased adoption and use of artificial intelligence (AI) technologies is changing the way industries operate. The construction industry is no exception, and neither is the practice of construction law. Construction teams are beginning to adapt to the use of AI tools, and construction lawyers are adapting as well. It is increasingly vital that people in both fields have an awareness of AI and how it is being integrated into the work of planning, developing, managing, and building projects. Awareness of AI’s limitations is similarly vital.

The term AI is a bit of catchall. In general, AI describes systems and software built to mimic human behavior and thinking by processing the types of inputs humans take in and producing human-like responses and outputs. The term as commonly used encompasses various related technologies: machine learning, neural networks, discriminative models, generative models, etc. The recent explosion in public awareness of AI has been spurred primarily by large language models (“LLMs”) like ChatGPT, which are “trained” on massive amounts of data and capable of generating responses to user prompts based on synthesis of that data. The rapid expansion of data storage and processing capacity has fueled these new technologies. The sudden publicity produced by ChatGPT has driven a feverish trend toward development (or marketing) of automation technologies and data processing as “AI.”

In the construction industry, humans have always been required to process and understand large amounts of text and data. They must read and understand dense and voluminous specifications, drawings, notes, schedules, and contracts, and apply them to the physical work of building. They make and keep large amounts of data such as site conditions, weather, manpower, schedule progress, and so on. Over the past two decades, there has been a push toward more data, not less, with building information modeling, computer assisted design, and other data-driven technologies. The varying degrees of human competence and diligence applied to these tasks and tools have led to many construction disputes and kept us construction lawyers plenty busy.

In recent years, contemporary generative AI technologies (i.e., those that function similarly to ChatGPT) have become very good at doing these sorts of tasks, quickly processing large amounts of text and answering questions about it. Unsurprisingly, companies have begun to create construction-specific applications. Generative AI programs like Civils.ai will process plans, site data, and contract documents, and allow users to ask questions about them. Programs like Buildots process schedule, productivity, and management information, and assist with work planning. Commonly used construction management platforms like Procore are incorporating AI-driven searches and automation.

So, with these sorts of powerful technologies, is AI going to eliminate mistakes, errors, and disputes on construction projects? Not likely. The first reason why lies in a well-known risk of generative AI: it “hallucinates.” AI hallucinations are responses that contain, or entirely consist of, false or incorrect information synthesized and fabricated from underlying data. They appear indistinguishable from factually correct responses. For example, in a notable case often discussed among lawyers, ChatGPT provided legal citations to entirely made-up authority in perfect citation format, which a lawyer then included in briefing to a court. The prospect of hallucinations means that a user should double-check an AI program’s responses before it trusts them.

The second reason lies in the way that the law of construction is intertwined with the documents, records, and data of construction. A contractor faced with a question regarding the drawings for a project might reasonably think, “Why should I submit a formal request for information to the architect and wait hours to days for a response when AI will give me the information in seconds?” The temptation to avoid a delay is a strong one, but a primary distinction is the legal risk shifting effect of processes like requests for information. A contractor who requests information from AI and relies on the answer may assume risk it would not otherwise have assumed had it requested the information from the architect. Coupled with the risk of AI hallucinations, this can create the potential for significant legal problems.

For these reasons, AI is not likely to truly upend construction methods and processes, at least in the near term. Neither is it likely to put construction lawyers out of business. Instead, lawyers need to be aware that designers, contractors, and consultants are incorporating AI tools into their workflows and that this trend is likely to accelerate as companies improve the technology. In litigation, traditional document discovery is poorly suited to reveal the inputs and responses of AI and other automated technologies, and lawyers should take this into account when assessing electronic discovery needs. A lawyer assessing a dispute will need to understand when, where, and how AI technologies were used on the project, what data and documents they were trained on or utilized as inputs, and whether generated responses can be accessed or exported. Lawyers also should consider and account for the use of AI and its attendant limitations and risks when negotiating and drafting contracts and evaluating project risk. It’s worth noting that lawyers are also starting to employ AI technologies of their own to streamline both discovery and contract drafting.

Used responsibly and knowledgably, AI technologies can efficiently harness the power of the copious data produced for and during construction projects. Contractors, designers, consultants, and owners, as well as their lawyers, should be aware of the technologies and how they can be used and misused. As more powerful and more accurate AI applications are developed, this will become even more essential. While we need not (yet) bow to our robot overlords, the use of AI in construction should no longer be ignored.