How will generative design impact the construction industry?

A recent WBEF webinar discussed the current and future applications of generative AI tools and their impact on jobs. Here are some of the considerations from our expert panel.

Application of generative tools

Leveraging data analytics and not imposing restrictions on the myriad of options generated can help identify ‘sweets spots’, which can then be translated into briefs for architects and engineers, explains Pim van Wylick, an Architect and Partner at PLANALOGIC. This can help stakeholders make informed decisions and can caution against venturing into unfavourable areas that could compromise the project due to regulations, sustainability, affordability or other factors, he says.

Generative scheduling is a parametric approach that uses algorithms to analyse options and explore new scenarios or (re)scheduling possibilities. Generative scheduling is applicable across the life cycle, not simply at the front end of a project or its execution, says Morgan Hays, Senior Vice President, Product, of ALICE Technologies. He cites various uses of the technology by his company including in the bid phase for tunnelling projects, in the detailed engineering phase for a phosphate mine and the feed phase for a lithium processing plant.

True generative AI and the ability to generate not only designs but also structural, mechanical, electrical and plumbing specifications from just a few parameters will revolutionise construction in the near future, believes Patrick Murphy, CEO and Founder of Togal.AI. Through the analysis of thousands of past builds and applying that learning to new projects, the predictive capability of generative AI to proactively provide key project requirements upfront is one of the most exciting elements of the technology, he says. He points out that a frequent problem in the industry is not making use of the wealth of historical data to inform new projects. Use of generative AI throughout the construction process can significantly reduce the likelihood of oversights such as a building specification not complying with disability requirements, incomplete plans, surprise change orders, scheduling delays and the apportioning of blame on construction sites, he says.

Laying the foundations

Getting the foundations of generative design set up correctly is an incredibly manual process, says Niknaz Aftahi, CEO & Founder of aec+tech. The AI is there to help and automate the process, but the design parameters and data inputs used by the algorithm need to be well defined, she says. Establishing an effective feedback loop is also central to a good foundation. By incorporating human judgement and evaluation, generic options can be refined into optimised solutions and the feedback provides for continuous improvement in the process, she says. As generative tools become more user-friendly and accessible, they are no longer restricted to a select group of technical experts, making collaborative decision-making increasingly important, she adds. AI can also enable better integration of multiple disciplines and a shift away from linear methods often used and that can result in having to revisit the design late in the development stage, says Pim van Wylick.

Risk mitigation

At its heart, generative scheduling is about risk mitigation, such as preventing oversights on projects or mitigating risk in the bid phase, believes Morgan Hays. He explains that without generative scheduling, bidding for example, is often based on previous approaches employed on projects. However, with the introduction of AI and generative scheduling, these constraints no longer exist as project stakeholders can explore alternative approaches to mitigate risks without having direct experience of them, he says.

The technology also plays a crucial role in facilitating higher-quality conversations among the general contractor, project coordinator and the owner. Without a generative scheduling tool, owners are at a disadvantage in such dialogues, lacking the means to effectively address concerns about project milestones and, for example, get a wayward project back on course, says Morgan Hays. An AI-based toolset creates opportunities for owners and their partners to engage in productive discussions, which can greatly enhance project outcomes, he says.

The three realities

Material substitution is a crucial aspect of decarbonisation assessments and making the right choices is essential, as even minor changes can have significant consequences, says Pim van Wylick. When considering materials, one major factor is synergy; generic designs are rarely monofunctional and often require a holistic approach, he explains. For example, aiming for a compact building based solely on its physical appearance may overlook other important metrics such as CO2 emissions. By analysing these trade-offs using generative tools, we can better understand the three realities at play:

• the physical reality of the building
• the financial reality of affordability and
• the sustainable reality of meeting regulations and certifications.

Implementation challenges

Lack of awareness and understanding of the technology is a major obstacle to the wider adoption of generative AI, says Niknaz Aftahi. This also makes it harder to establish a business case for the required investments in software, hardware, training and infrastructure, she says. Additionally, the success of AI models relies on vast amounts of data, but accessing and sharing big data continues to be a challenge for the construction industry. Lastly, she notes, there are regulatory and legal concerns surrounding the use of generative AI, for example, intellectual property rights of others when creating visualisations or renderings, adding another layer of complexity and compliance.

Black box algorithms, where the AI’s thought process is hidden, making it impossible to see how decisions were reached, should be avoided says Pim van Wylick. A strength of a well-implemented rule-based AI system is the ability to check outcomes, reducing repetitive elements and rework for architects and allowing them to focus their core skills, he says.

Impact on jobs

Generative AI can improve intuitiveness and usability of software tools like BIM. Currently, the extent and effectiveness of the implementation of BIM varies. Patrick Murphy anticipates that in the future, generative AI will make BIM more accessible to a broader range of skill sets, including more workers on-site.

Advances in AI technology will increasingly allow humans to focus on higher value tasks, freeing us from the mundane and remedial work that consumes a significant portion of our lives, says Patrick Murphy. Activities such as checking various codes and taking measurements can be effectively delegated to computers, allowing humans to concentrate on creativity, he says. Although the advancement of AI and its impact on jobs has raised concerns, it is important to consider how technology has created many of today’s jobs, he says.

The use of AI can heighten creativity by ensuring each project yields a unique outcome, presenting the best overall solution tailored to specific criteria, says Pim van Wylick. This will empower architects to act as editors-in-chief, curating and filtering the options that align with their desired themes or objectives, he says.

Adopting a more knowledge-based approach to construction through generative design can reduce routine and repetition in the design process, more easily allow for customer configuration, and increase speed of delivery. How can the advantages of generative design be best applied to help meet demand for infrastructure and housing for example, cut costs and improve productivity and efficiency. What are the challenges implementing a generative design approach and how can they be overcome? How is generative design being used today and how will it change the way we build tomorrow?