Of all the Industry 4.0 innovations transforming the business world, construction and architecture may benefit the most from machine learning (ML) and artificial intelligence (AI).

New machine learning-powered tools are helping support architects and designers by drawing on vast amounts of architectural data to provide insights and automate repetitive design tasks.

These three use cases of ML in architecture show how the field is already using the technology and how it may innovate further over the next few years.

1. ML in Design Automation

For architects, one of the most significant advantages of machine learning may be its ability to perform repetitive tasks that are conventionally difficult to automate. These tasks are time-consuming and repetitive but are just complex enough to require human problem-solving capabilities, rendering them too complicated for simpler tools like robotic process automation (RPA) to manage on their own.

AI and ML, however, can use pre-existing data to automate these slightly-too-complex architecture tasks. These tools enable design strategies that take advantage of existing data to generate entirely new architectural information.

One example of new task-automating ML in practice is the AI algorithm Finch, a design feasibility tool that automatically generates spatial configurations according to predetermined parameters as the architect adjusts the dimensions of the total space — an example of what’s called parametric design.

As the architect adjusts the length and width of the space, Finch automatically adjusts its design by shifting furniture and resizing rooms as necessary to ensure all the required elements fit inside the space.

The tool’s designers intend architects to use Finch to define zones during the initial stages of a project. These zones can be refined as needed according to the specific requirements of the assignment.

2. AI and ML in Sustainable Design and Operations

Sustainability has quickly emerged as a critical topic of discussion in the architecture world. Watchdogs estimate that around 40% of total global carbon emissions can be traced back to construction and building operations. Building structures that require less energy to operate and maintain could allow the industry to significantly reduce global emissions. The right algorithm could help managers run their buildings more efficiently — reducing the carbon cost of building operations.

AI-powered building energy analytics, for example, can be deployed to automatically adjust building temperature based on factors like occupancy, weather, and other system parameters. These systems can integrate with digital construction tools, providing them with data from before the building was even in operation.

Networked devices, like smart vents and air quality monitors, can make the system more efficient and provide building managers with real-time feedback on the performance of HVAC and other essential building systems.

These tools can also help the HVAC system coordinate with other building systems. A great example is one IBM project, the Al-Bahar Towers in Abu Dhabi, which features louvers that building managers can articulate open or closed as necessary, helping to reduce solar gain and cut carbon emissions.

Some AI and ML algorithms could even help track the performance and health of individual components. For example, gearboxes are a necessary component for exterior building clocks. When a gearbox breaks down, the clock arms will stop moving and the face will display an inaccurate time.

Computer scientists have already designed algorithms capable of monitoring gearbox performance and health in real-time. This allows building operators to catch gearbox failures before they happen and improve gearbox fault analysis. Both of these tools could help building managers more effectively prevent and respond to specific types of clock failures.

3. ML for Generative Design

In some cases, ML tools may also enable the design of unique structures that may have been impossible or impractical to design with a conventional approach. These tools can use previous architectural projects to generate entirely new designs, allowing architects to explore design techniques that they may have never uncovered on their own.

Generative design is becoming increasingly popular in architecture, engineering, and design. With this approach, an AI algorithm trained on vast amounts of architectural information generates new designs from scratch.

With each iteration, the algorithm reinforces its ability to generate these designs, helping it move closer and closer to a practical solution to a given problem — like a new motorcycle swingarm or bridge design.

The architect, engineer, or designer takes on a curatorial role by selecting interesting designs, ruling out the ridiculous or impossible, and guiding the AI towards potential answers to an assignment’s goals.

These generative design tools are often deployed in combination with other novel technologies, like 3D printers. A 3D printer is capable of manufacturing the abstract or unconventional designs that AI generates.

How AI and ML May Help Transform Architecture

New AI and ML tools are helping architects work faster, design more sustainable buildings, and create designs they may not have been able to create on their own. These tools are likely to become more sophisticated over the next few years as architects continue to experiment with them and adopt innovations from AI research.

cover image © pexels