AI is at its best when it is used to enhance productivity and improve the lives of those it affects. When used correctly, AI can also save lives.
That’s the vision driving a new project at SAS, where applied AI models and cameras create a simulated work environment focused on worker safety.
This project is more than just a proof of concept; it’s a real-world application of AI designed to make manufacturing and industrial environments safer for workers. And with workplace injuries on the rise, the stakes couldn’t be higher.
Luckily for you, the reader, I’ve got the inside scoop on how this vision is being turned into reality and why it matters now more than ever.
The vision behind the project
Udo Sglavo, Vice President, Applied AI and Modeling, who leads the newly formed Applied AI and Modeling R&D division at SAS, explains that the project aims to “productionize services” developed and refined over years of consulting and fieldwork.
“One of the models we’re working on aims to uncover safety gaps and identify potential risks to workers in real-time, helping reduce workplace injuries and property damage,” Sglavo explains.
With workplace injuries and fatalities rising—up 5.7% in 2022, according to the U.S. Bureau of Labor Statistics—there’s a pressing need for better safety measures. This is where AI steps in.
Sglavo’s team, in collaboration with SAS’ IoT division and IT consulting and digital services provider SoftServe, has been working on a solution aptly named SAS for Worker Safety. The idea? Use AI to monitor compliance with safety regulations in real time, reducing risks before they become tragedies.
Creating a simulated environment
To build an effective AI model, you need data—lots of it. But how do you gather data in a way that doesn’t depend on real-world customers? “One way is to do this completely artificially, using synthetic data and digital environments, but nothing is as good as the real thing,” Sglavo notes.
That’s when one of his teammates came up with a brilliant idea: why not create a simulated work environment right in our workshop? I mean, we’ve already seen the magic that happens behind the scenes in SAS’ massive warehouse workshop.
This setup would allow research and development teams to collect the data needed to enhance their models without putting any actual workers at risk.
“This camera and this environment allow us the flexibility to design a number of these safe and unsafe incidents so that we can rapidly prototype and fine-tune our models for our customers,” Sglavo adds.
This innovative approach has turned the workshop into a dynamic testing ground where AI can be trained in real-world scenarios. The flexibility of this simulated environment means the team can generate countless variations of safety incidents, refining the AI’s ability to detect and respond to potential hazards.
Charting this project's real-world impact
The results so far have been promising. Varunraj Valsaraj, Sr. Manager, Applied AI and Modeling, who has been deeply involved in the project, says, “We have started testing this worker safety solution in our customer warehouses, and we are seeing some incredible results. For instance, these models are able to detect 100% of the unsafe incidents.”
This means that organizations can implement AI-driven safety measures without needing to be experts in AI themselves. The models can be integrated into existing shop floor operations, analyzing real-time streaming data and alerting managers to potential safety violations before they lead to accidents.
Why worker safety matters
The importance of worker safety cannot be overstated. Every year, thousands of workers are injured or killed on the job, often due to preventable issues. By using AI to monitor safety compliance, organizations can protect their workers and reduce costs associated with workplace injuries, such as medical expenses, legal fees and lost productivity.
Also, compliance with regulators like the Occupational Safety and Health Administration (OSHA) is critical for avoiding fines and maintaining a good reputation. AI can help ensure that organizations adhere to these regulations, providing peace of mind for both employers and employees.
Beyond immediate financial and legal implications, there’s a broader cultural shift toward prioritizing worker well-being, and SAS' innovative approach can be at the forefront of this movement.
Looking ahead
This project is still in its early stages, but its potential to revolutionize workplace safety is immense. By creating a simulated work environment with AI, SAS sets the stage for a future where safety is monitored and proactively managed in real-time. The early results during testing show that this technology could drastically change how organizations approach safety.
As the project progresses, we may learn that AI models are completely necessary for industrial operations if organizations want to reduce accidents to a rare occurrence. The journey is only just beginning.
Go further: 5 key questions to ask when developing trustworthy AI models
WANT MORE GREAT INSIGHTS MONTHLY? | SUBSCRIBE TO THE SAS INSIGHTS NEWSLETTER
