Companies look to AI for smarter approach to workplace safety

Artificial intelligence tools are changing dramatically the way people work and in many cases making workers’ jobs easier and streamlining processes.

So it’s no wonder companies are looking toward AI to make their workplaces safer.

While these tools aren’t meant to replace safety professionals, they can help make sense of data and trends to help prevent future and also provide an extra set of “eyes” to detect safety infractions, experts say.

“I think AI really offers the opportunity for organizations to focus on leveraging their large volumes of data to drive predictive insights, instead of relying on lagging indicators after employees are already harmed,” says Sarah Ischer, a senior program manager at the National Safety Council, a nonprofit safety advocacy organization. The NSC recently released a white paper on how employers can apply AI technology to reduce injuries.

Data and AI offer a variety of applications in the workplace, according to NSC’s report, including:

• the use of cameras and wearable sensors to collect data to determine posture and motion during lifting and other physical activities, and provide feedback
• cameras with computer vision to monitor and detect PPE noncompliance, spills, fires and poor housekeeping
• using AI tools to streamline reporting and compliance as well as analytics

Right now it’s mainly being used by larger worksites, but it is “growing in smaller and medium-sized companies as the software and tools become more affordable,” Ischer says.

Some of the costs can still be prohibitive depending on the application or use, she says.

Ischer advises companies to start small and understand specifically what you really need from the tools because there are many different options and costs.

Charles Hunt, chief operating officer at Massapequa Park-based Able Safety Consulting LLC, which provides OSHA compliance assistance and training, agrees, noting some of the costs can be high, especially for live video feeds.

For the applications he uses, it costs about $1 per 1,000 images and for live video it costs 1 cent per second. On its own that doesn’t seem like a lot, but imagine a worksite that needs video on for 8 to 12 hours, says Hunt, also president of Lergent Corp., a software developer for e-learning.

He uses AI as part of facial recognition software his firm developed for logging into safety courses to make sure the person logging in is who they say they are.

Hunt is seeing AI used at worksites to examine proximity and motion.

For example, it can detect that you have 10 workers on one floor and 15 on another so you can account for everyone. Hunt has also seen it used in confined spaces to make sure a worker is still moving and not in physical trouble, to alert a manager if a worker is coming too close to an edge that could risk a fall and to gauge unsafe proximity to a forklift or other machinery.

It has wide capabilities, but it’s not meant to replace safety professionals, says Holly Pups, an environmental health and safety editor at J.J. Keller & Associates Inc., a firm that provides safety and compliance solutions for employers across the United States.

“AI complements the work of a safety manager who understands the human side, where AI can only see the datapoints which it’s programmed to see,” she says.

Stefan Borovina, a partner at Goldberg Segalla in Garden City, whose practice focuses on OSHA and workplace safety, notes that generally speaking, a safety professional can only be in one place at a time.

They can’t monitor multiple floors of activity in real time at all times and that’s where AI can help, he says. 

But be clear, “it's not going to replace the employer's obligation to keep the workplace free from hazards.,” Borovina says.

Pups advises companies to evaluate existing systems and see how they can integrate AI within those existing systems.

Borovina agrees, advising to try to home in on one particular safety issue and see how AI can help.

Also, try to get employee buy-in and feedback early on, Pups says.

“There’s a 'Big Brother' fear,” she notes.

Companies should explain to employees why they’re implementing these tools and how they can help them, she says. Work with employees throughout planning and implementation to increase their understanding and buy-in, Pups says.

David Lemon is senior customer success manager at intenseye, an AI-powered workplace safety platform. Using existing facility cameras, intenseye identifies and analyzes over 40 unsafe behaviors and conditions happening at a facility, like ergonomics deviations, leakage and spills.

With the platform, there’s “facial blurring” to protect the worker, Lemon says.

“The anonymization strengthens the data set,” he says.

It’s not about blaming a particular person, “but rather identifying overall unsafe trends and behaviors,” Lemon says, noting the company's seen exponential growth in intenseye's use across multiple industries since the firm's inception in 2018.