Share

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus convallis sem tellus, vitae egestas felis vestibule ut.

Error message details.

Reuse Permissions

Request permission to republish or redistribute SHRM content and materials.

Companies Turn to AI to Improve Workplace Safety


Two warehouse workers walking through a warehouse.


​While the use of generative AI to save HR professionals time and boost their productivity has captured headlines, another use of artificial intelligence in the workplace is having a significant impact: improving worker safety and reducing on-the-job injury and fatality rates. 

Workplace fatality rates rose significantly in 2021, according to data from the U.S. Bureau of Labor Statistics (BLS), to the highest annual rate since 2016.

To help reverse the trend and enhance occupational safety, more companies are turning to AI, advanced data analytics tools and other safety-related technologies, according to a report from the National Safety Council (NSC), a nonprofit safety advocacy organization in Itasca, Ill. The report evaluated recent findings from academic and industry journals and identified three new technologies having a strong impact on preventing workplace illness, injury and death.

SHRM Online spoke with Sarah Ischer, a senior program manager on the NSC's workplace safety team, to go in-depth on how these technologies can help HR and environmental, health and safety (EHS) professionals make the workplace safer; explore specific use cases of the tools; and examine barriers that keep companies from adopting such technologies.

Three Technologies Lead the Way

The three technologies highlighted in the NSC report not only can improve safety, but if used effectively they can also save organizations time and money. The technologies are as follows:

Natural Language Processing

Commonly used in other areas of HR, such as for sentiment analysis of engagement surveys, natural language processing (NLP) technology is now being used to reduce laborious manual processes and streamline safety reporting.

EHS and HR professionals typically have to manage and analyze large amounts of safety-related data in the form of written reports, videos, images, statistics and more. NLP technology can rapidly analyze such data and extract key findings and insights, Ischer said.

The Los Alamos National Laboratory is among organizations using NLP and related analytics tools for such purposes. The laboratory partnered with Toronto-based vendor Cority to help improve its data quality and safety culture. As detailed in the NSC report, the laboratory's existing data management system struggled to reliably track data, resulting in difficulties reporting or checking on EHS indicators.

By using Cority's cloud-based platform, the laboratory was able to decommission more than 75 software applications and unify EHS management, employee wellness and operational reporting with a single-source-of truth platform. The technology now automatically generates dashboards featuring advanced trend analytics and safety reporting.

Computer Vision Technology

Computer vision technology monitors video footage and images in the workplace, detecting a wide variety of objects to identify risks. "It includes things like automated alerts for equipment malfunctions, understanding if someone is wearing the appropriate personal protective equipment [PPE] or even identifying actions that can lead to workplace violence," Ischer said.

For example, an organization might use next-generation software in a CCTV video management system to monitor activity in a warehouse or a manufacturing environment. "The software essentially is trained to know what to look for," Ischer said. "If you're looking at use of PPE, for example, when someone walks through a frame you can see whether they're wearing a hard hat or not. If they're not it would trigger automated alerts that would let managers or safety professionals know that PPE isn't being properly used."

Ischer said new AI tools used with computer vision technology can analyze historical data to identify trends and patterns. "You might see that in a specific warehouse or during a specific shift that PPE is only being used 50 percent of the time, which allows an organization to follow up to understand the reasons why," Ischer said.

Predictive and Prescriptive Analytics Engines

Predictive and prescriptive analytics use AI to help learn cause and effect from historical safety data, with the goal of helping organizations avoid safety incidents before they occur—or to mitigate the fallout once such events have happened.

Next-generation AI can analyze large datasets of historical sequences of events and examples of successful mitigation strategies to detect potential risks, the NSC report found.

"It allows systems to be configured to deliver alerts or recommendations when a problem occurs to avoid risks connected to the event," Ischer said. For example, the technology could alert managers to a spill in a warehouse to ensure rapid cleaning and to keep workers from walking through the area.

"You might also have a camera focused on a tank or pipe that measures the pressure coming in or out," Ischer said. "That assesses the odds of a potential hazardous release or spill to avoid problems."

Such technology also can help mitigate incidents of workplace violence when integrated directly with facility controls and permit management systems. For example, some tools can activate facility lockdown procedures if a potential workplace violence incident is detected, according to the NSC report.

Costs and Other Barriers to Technology Adoption

The NSC report found that while the cost of the three highlighted technologies can vary widely based on the type used, they aren't out of reach for smaller and midsize organizations.

Some NLP models including ChatGPT can be accessed through pay-as-you-go application programming interfaces and publicly available, pre-trained AI models that can be fine-tuned on smaller datasets, the NSC report found. "That allows smaller organizations to leverage their capabilities in specific use cases and allow continuous retraining in the field," authors of the NSC report wrote.

"Cost can be a barrier for some organizations, but it does depend on what a company needs, its workplace safety objectives and what it already has in place," Ischer said. "For example, if they already have cameras onsite, the cost of adding new software for computer vision safety purposes can be manageable."

HR and safety leaders also need to consider data privacy and protection issues connected to using these technologies, Ischer said. "Employers should be upfront with employees about what data they'll be collecting with these tools and how they intend to use it," she said. "That's important in getting employee buy-in to using the data to improve workplace safety."

Dave Zielinski is principal of Skiwood Communications, a business writing and editing company in Minneapolis.

Advertisement

​An organization run by AI is not a futuristic concept. Such technology is already a part of many workplaces and will continue to shape the labor market and HR. Here's how employers and employees can successfully manage generative AI and other AI-powered systems.

Advertisement