A Boston-based construction company has been developing an artificial intelligence (AI)-based system for predicting construction site injuries.

With construction site fatality rates roughly five times higher than any other industry, construction sites are considered among the most dangerous places to work. As such, Boston construction company Suffolk, along with computer vision company SmartVid, have been working on potentially life-saving technology that predicts when construction site accidents are likely to occur. Suffolk and SmartVid have been training deep learning algorithms on construction site images and accident reports so that the AI system can monitor construction sites and flag scenarios that hint at possible accidents. Such hints might include workers not wearing gloves or working too closely to dangerous machinery.

Recognizing that one company would not be able to offer enough data to efficiently train the algorithms, Suffolk invited a number of its construction company competitors to join a consortium called the Predictive Analytics Strategic Council to contribute data to improve the system’s performance.

In addition to potentially saving lives, the project also demonstrates the potential for AI-enabled computer vision to monitor and predict activity in the workplace, thereby potentially improving productivity and reducing significant cost overruns, particularly in the construction industry. However, the technology is also a demonstration of the future workplace where AI algorithms will monitor, quantify and optimize employees' work.

The work was presented by Jit Kee Chin, chief data officer and an executive vice president at Suffolk, at the EmTech Next conference held June 11 and 12. The conference is hosted by MIT Technology Review and it examines global trends in the future workplace.

Advances in technology are rapidly changing the face of the construction industry with one recent report estimating that more than 7,000 robots will be working in the construction industry by 2025. Similarly, sensors are being used to monitor structures such as bridges for performance under a variety of conditions, informing builders when repair or replacement is necessary.

To contact the author of this article, email mdonlon@globalspec.com