AI Regulation in Construction and BIM

This page is a translation of the original Portuguese text, provided for convenience. It deals with Brazilian law. If there is any discrepancy, the Portuguese version prevails. Read the original in Portuguese

Artificial intelligence is transforming construction through BIM, but the lack of specific regulation creates legal risks that demand immediate attention from builders and developers.

The Rise of Artificial Intelligence in Construction and BIM

The construction industry is undergoing an unprecedented technological shift. Integrating artificial intelligence (AI) tools with Building Information Modeling (BIM) now allows entire projects to be optimized automatically, from structural calculations to job site logistics planning. Machine learning algorithms analyze thousands of variables at once to suggest more efficient materials, predict schedule delays, and identify clashes between design disciplines before construction even begins.

From a legal perspective, Brazilian law still has no specific legislation regulating the use of AI in construction. Bill No. 2.338/2023, which would establish the AI Legal Framework in Brazil, addresses the subject in general terms and does not deal with the particulars of the construction sector. This leaves significant gaps, especially regarding civil liability for automated decisions that affect building safety.

By its collaborative nature, BIM involves multiple parties (architects, engineers, builders, suppliers) who all feed a shared digital model. Once AI layers are added to the process, complex questions arise. Who is responsible for a design error suggested by an algorithm? Can the builder that adopted the automated recommendation claim a defense that excludes its liability? Brazilian legal scholarship and case law have not yet settled on answers to these questions.

Civil Liability and Automated Decisions in Engineering Projects

One of the most sensitive points where AI, BIM, and construction meet is assigning responsibility when automated decisions lead to construction defects or safety failures. The Brazilian Civil Code, in Articles 927 et seq., establishes a duty to compensate for damage caused by an unlawful act and, in certain situations, adopts strict liability. The Consumer Protection Code, which applies to relationships between developers and property buyers, provides for the strict liability of the supplier for product defects.

Suppose an AI system integrated with BIM suggests replacing a structural material with a cheaper one, and that substitution compromises the integrity of the building. The chain of liability then becomes complicated. At least three parties may be involved: the developer of the AI software, the designer who validated (or failed to validate) the automated suggestion, and the builder who carried out the project. The trend we see in comparative legal scholarship is that human oversight remains the central element of accountability. In other words, the licensed professional who signs off on the project is jointly and severally liable, even if an algorithm originally proposed the decision.

The issue becomes even more delicate when we consider that many BIM programs with built-in AI operate as “black boxes,” whose decision-making processes are not fully transparent. The principle of explainability, set out in the AI Legal Framework under discussion in Congress, would require systems used in engineering projects to be able to show, in understandable terms, the reasons behind each automated recommendation. In practice, not every system available on the market meets this requirement.

Qualified human oversight remains indispensable: no algorithm replaces the technical responsibility of the engineer or architect who signs the project.

Technical standard NBR 15.575 (issued by Brazil’s technical standards body), which sets performance requirements for residential buildings, makes no mention of automated tools in project design. Because of this regulatory gap, professionals and companies in the sector need to adopt their own human validation protocols for every critical decision suggested by AI systems. Otherwise, they bear the full risk of any resulting failures.

Data Protection and Intellectual Property in AI-Powered BIM Models

BIM models powered by artificial intelligence process large volumes of data. These include proprietary technical information, material performance data, supplier specifications, and, in some cases, personal data of workers and residents. This puts Brazil’s General Data Protection Law (Law No. 13.709/2018), known as the LGPD, at the center of the discussion on digital governance in construction.

In projects that use collaborative cloud-based BIM, sensitive data moves across multiple servers and jurisdictions. Builders that use international BIM platforms with AI must make sure their processing of personal data complies with the LGPD. That includes having appropriate legal bases, preparing data protection impact assessments (RIPD), and observing the principles of purpose, adequacy, and necessity.

Equally important questions arise in intellectual property. Can an architectural design generated or substantially modified by artificial intelligence be protected by copyright? Law No. 9.610/1998 (the Copyright Law) protects creations of the human mind, which raises doubts about who owns design elements generated by algorithms. The position we observe in Brazilian legal scholarship is that works generated exclusively by AI, without substantial human creative input, do not enjoy copyright protection. However, projects in which AI acts as an auxiliary tool under the creative direction of a professional remain protected.

Another aspect that deserves attention is trade secrecy. BIM models contain strategic information about construction methods, costs, and proprietary technical solutions. When these models are processed by cloud-based AI systems, there is a risk that commercially sensitive data will be used to train algorithms that benefit competitors. Licensing agreements for BIM software with AI should include specific clauses on the use of data entered by users, expressly limiting its use for training machine learning models.

Building Safety and the Duty of Diligence When Adopting AI

The structural safety of buildings is a matter of public order, governed by the Civil Code, by technical standards of the ABNT (Brazilian Association of Technical Standards), and by specific municipal legislation. Introducing AI systems into project design and construction does not change this imperative: a building must be safe regardless of the technology used to conceive it.

We note that using AI for structural sizing, load analysis, and foundation planning offers significant benefits in precision and efficiency. Those benefits materialize only when the systems are used within appropriate parameters and under competent technical supervision. Internationally documented cases show that structural optimization algorithms can, in certain circumstances, suggest solutions that satisfy the mathematical parameters of the problem but ignore essential practical variables, such as local geological conditions, the history of how the land has been used, or regional climate particularities.

The duty of diligence requires builders and designers to critically verify every AI-generated recommendation, especially those that affect structural safety. That verification must be documented, creating an auditable record showing that a licensed professional reviewed each automated decision before it was implemented. Without this documentation, a company’s position in a lawsuit can be significantly worse.

Professional liability insurance is another issue to consider. Insurers are beginning to include specific clauses on the use of AI tools in engineering projects. Professionals and companies using these technologies should review their policies to make sure the coverage includes risks associated with automated decisions, avoiding unpleasant surprises if a claim arises.

Regulatory Outlook and Best Practices for the Sector

The regulatory landscape for artificial intelligence in Brazil is still taking shape. The AI Legal Framework, which is still moving through the legislative process, is expected to establish general principles applicable to all sectors, including construction. Among the principles most relevant to the construction sector are algorithmic transparency, meaningful human oversight, non-discrimination, and accountability.

At the regulatory level below statutes, the Federal Council of Engineering and Agronomy (CONFEA) and the Council of Architecture and Urbanism (CAU) have not yet issued specific resolutions on the use of AI in projects. Even so, the current rules on technical responsibility (such as the Anotação de Responsabilidade Técnica (Technical Responsibility Annotation) and the Registro de Responsabilidade Técnica (Technical Responsibility Registration)) already make the professional fully responsible for the project, regardless of the tools used to prepare it. In practice, this means an engineer or architect who uses AI in their projects remains the professional responsible before the professional councils and before third parties.

Given this landscape, we recommend that construction companies that use or plan to use AI integrated with BIM adopt a set of best practices. First, it is essential to implement internal AI governance policies that clearly define the roles and responsibilities of each party in the project workflow. Second, any automated decision affecting structural safety, choice of materials, or sizing should go through documented human validation. Third, contracts with software vendors should include clauses on confidentiality, limits on data use, and liability for algorithmic failures. Fourth, the company should keep detailed records of all interactions with AI systems during the design process, creating a complete audit trail.

We also note that Decree No. 11.856/2023, which instituted the Brazilian Artificial Intelligence Plan, reinforces the need for strategic sectors such as construction to develop their own guidelines for the responsible use of the technology. Getting ahead of these regulatory requirements is not just a matter of compliance. It is also a competitive advantage for companies that want to position themselves as leaders in responsible innovation.

This subject is directly related to artificial intelligence regulation, which we cover in a dedicated article.

This subject is directly related to LGPD civil liability for data leaks, which we cover in a dedicated article.

As informações deste artigo são de caráter informativo e não substituem consulta jurídica individualizada. Cada caso possui particularidades que exigem análise específica.