Encouraging responsible AI in the construction sector

Leaders in the construction sector, like many other sectors, have high expectations that AI technologies can help to improve their work. AI may be crucial for the future of the sector, enabling cost-saving methods and contributing to sustainability goals like reducing carbon emissions and air pollution. However, there is also anxiety about AI. A study by IFS reveals that a third of UK respondents are anxious about workforce readiness and the quality of resources for learning about AI. Randstad reports that a quarter of UK construction sector workers would rather change jobs than learn to use AI. To enjoy the benefits of using AI technologies in this sector while minimising negative effects on workers and others, investment is needed for education and training for construction sector workers on techniques for responsible AI development and management. This article will review applications of AI for the construction sector (for environmental impact, workplace safety, productivity and design). Then we will review barriers to the adoption of AI in the sector, and the training topics that can reduce those barriers. 

Environmental impact 

The construction sector is heterogenous, containing diverse subsectors including building contractors, architecture and design, infrastructure, landscaping and related services. Construction sector activities have big effects on people’s lives in the UK. More than five in six people live in urban environments. The UK aims to achieve net zero carbon emissions by the middle of this century, and buildings are responsible for around 17% of all UK emissions. AI technologies can be powerful tools for achieving net zero goals. AI tools can be used for carbon reduction and more rapid, detailed and accurate modelling of environmental impact assessments. There are AI tools for building management, using sensors and historical data to regulate heat, minimise energy consumption and predict maintenance needs. On a larger scale, AI technologies have been developed for ‘smart cities’ projects, such as smart grids and virtual power plants, that are expected to reduce energy consumption and carbon emissions. 

Workplace safety 

Various AI tools have been developed to improve safety for workers at construction sites. These include computer vision tools that automatically monitor hazards, safety equipment and compliance with safety protocols. ‘Predictive maintenance’ systems use historical data to predict breakdowns and maintenance needs before problems arise. AI can also be used to detect defects in building materials and concrete. 

Productivity and project planning 

Productivity growth within the construction industry has been slow. Global productivity within the industry has only increased by 1% annually over the last two decades. Between 2014 and 2019, 26 billion USD was invested into engineering and construction, but the investments did not lead to a change of core construction methods. AI may enable more productivity within the construction sector. New tools promise to streamline project planning by automating the optimisation of schedules, supply chains, and other resource management tasks. Other tools automate compliance checking tasks, improve accuracy of specifications, or save labour on other tedious or time-consuming administrative tasks. 

AI adoption in design and architecture 

So far, rates of AI adoption vary a lot across the construction sector. The Royal Institute of British Architects (RIBA) mentions in its 2024 AI Report that 41% of architectural practices that responded to their survey have used AI at least occasionally. However, only a small minority of these practices use it for all (2%) or most projects (4%). The uptake of the use of AI within this subsector of construction has been driven by readily available and easy to use consumer-facing generative AI platforms such as ChatGPT. 

According to the RIBA report, AI tools are used most often during the early design phase, for visualisations or to rapidly produce ideas and rough plans based on various specifications. Increasingly, AI tools are used to optimise floor plans and to augment other parts of parametric design processes. However, other potential applications are less common. For example, many architects expect AI to be helpful in modelling and simulations, but AI is rarely used for these tasks. RIBA attributes this hesitation to a lack of mature tools and training and skills for architects. Architects lack formal training in the responsible use of AI which is hindering its uptake. 

Breaking down barriers through education 

Uptake of AI tools within construction remains low compared to other sectors. Some barriers include: 

• Lack of skills: There is a shortage of AI developers with the necessary skills to develop solutions targeted to the construction sector. In addition, workers in the construction sector lack formal training in AI.
• Novelty: During a build, small mistakes can have significant negative implications, both financially and in terms of human health and safety. Builders tend to trust proven construction methods above newer, untested ones. In addition, architects and other construction sector workers have concerns about how to use AI tools ethically and responsibly. 
• Investment: Another factor explaining low uptake is that some AI-powered technology, such as robotics, have high initial costs. At the moment, many small companies within the construction industry do not see a business case for investing in such technologies. 
• Data availability: While many architectural practices might have access to a wealth of well-structured data, many construction companies do not. Building contractors are often small companies that generate little data. Other companies may have large amounts of data that are unstructured or siloed, making them difficult to exploit in data-driven tools like AI. For the potential of AI to materialise, these companies need to work together but their data are often not interoperable. Regulatory challenges can form a barrier in relation to the use of AI with construction including liability, copyright, and intellectual property. 
• Workplace impacts: Many construction sector employees are anxious about the effect that AI technologies may have on their job security or on the ways that they work. 

These barriers, however, can be addressed in part by providing more accessible education and training on topics related to responsible AI. 

• Data fundamentals: Managers and other leaders in the construction sector can improve their understanding of new technologies and regulations without needing to become data scientists themselves. If construction sector leaders have a strong working knowledge of data fundamentals, such as what can and cannot be done with data and how to process data responsibly, then they will be better positioned to identify opportunities that suit their circumstances. 
• Risk management: Risk management techniques can help construction firms to identify and mitigate risks, including risks relating to the deployment of novel technologies. Risk management should be a crucial element in the responsible adoption of AI. 
• Business case development: Decisions about investment in new technologies can be handled with an informed approach that maps business challenges to the right technologies. Construction sector leaders—especially those from smaller companies—might benefit from a systematic approach to appraising costs and benefits. Practical training can help leaders to see how they can consider scalability and involve their stakeholders. 
• Workplace education: Responsible AI training at all levels of the workforce can help prepare employees to use AI technologies safely and responsibly. This education can also help them to more clearly evaluate their circumstances and advocate for their interests. 

At Trilateral Research, we are excited to be embarking on a project, together with UKRI and Innovate UK, to educate AI leaders and to promote responsible AI to the UK construction sector. We want to help bring about a cleaner, safer and more just future in our built environment. 

Our Ethics Innovation Team has extensive experience developing AI training and educational resources tailored for diverse kinds of companies and organisations. We create bespoke courses tailored to the strategic interests of our clients. Our courses are therefore never the same, but always adjusted to the skill levels and learning needs of the people we reach. If you are interested in personalised responsible AI training, please contact us by email dcs@trilateralresearch.com to discuss your requirements. Our team would be happy to help.