The artificial intelligence field dates to 1956 and the Dartmouth Summer Research Project on Artificial Intelligence. It’s taken more than 65 years, but today the occupational safety and health (OHS) field is inundated with news, articles, webinars, workshops, product PR, podcasts, a global AI and safety summit, and numerous U.S. conference presentations and panel discussions on the impact of AI on OHS. 

Look no further than this year’s American Society of Safety Professional’s (ASSP) national meeting in Denver (August 7-9) where AI is the topic of at least 11 educational sessions: 

• “Artificial Intelligence in EHS: Conquering fears and embracing frontier;”
• “Revolutionize Ergonomic Assessments & Training with Artificial Intelligence;”
• “Pros and Cons of an Artificial Intelligence-Driven Future;”
• “How Artificial Intelligence Enables Proactive Workflows to Improve OHS Management Systems;”
• “How Can Artificial Intelligence Improve Workplace Safety?;”
• “Emerging Trends of Artificial Intelligence in ESH;”
• “Applying Artificial Intelligence & Machine Learning;”
• “From Current State to Artificial Intelligence & the Next Frontier;”
• “Transforming ESH: Apply Artificial Intelligence & Machine Learning;”
• “Artificial Intelligence as a Safety Partner;” 
• and “Artificial Intelligence and Wearables.” 

The obvious potential

According to none other than ChatGPT, when asked, “Artificial intelligence (AI) has the potential to greatly enhance occupational safety by identifying risks, predicting accidents, and even suggesting preventive measures.” I wager this is a safe bet. 

Greatly enhance? Here are some of the benefits: 

1. AI can use advanced algorithms to identify potential safety hazards through pattern identification — an application already used in cameras and drones. 

2. AI can be used (though some pros are hesitant and there are currently flaws) to create safety policy documentation. 

3. It can potentially revolutionize training through AI-generated videos, written prompts, personalized learning and identifying learning gaps through data analysis and metrics to create profiles based on these gaps to be fill with appropriate training. 

4. AI allows real-time data collection to monitor workers and identify exposure risk. 

5. Environmental sensors and biosensors collect data on biological, physical or chemical changes that can flag early warning signs of occupational disease or distress. 

6. And, of course, there are robotics that rely on AI sensors to remove human contact from risky work tasks.

Who benefits?

Here's a critical, ethical issue recently addressed in the paper, “Occupational Safety and Health Equity Impacts of Artificial Intelligence: A Scoping Review,” published in 2023 in the International Journal of Environmental Research Public Health. The issue: Differences in work-related injury and illness incidence, morbidity and mortality are closely linked with social, economic and environmental disadvantage, according to the authors of the scoping review. 

Research suggests that wages, education, race, and place of birth are all associated with OSH mortality and job insecurity. Studies suggest that workers from certain racialized ethnic minority groups, such as Black, American Indian/Alaska Native, and Hispanic, those possessing a high school degree or less, or are foreign-born, are at a disproportionate risk for negative OSH outcomes.

A tragic reality

For confirmation, check the six workers killed in the Baltimore Francis Scott Key Bridge collapse this past March: Maynor Yassir Suazo Sandoval, Miguel Angel Luna Gonzalez, Jose Mynor Lopez, Alejandro Hernandez Fuentes, Dorlian Ronial Castillo Cabrera, and Carlos Daniel Hernández. All were from El Salvador, Guatemala, Honduras or Mexico. They all worked for an independent contracting company.

In an edition of Health Affairs, USC researchers found that men ages 18 to 64 who are Latino immigrants have the highest average workplace injury rate at 13.7 per 1,000 workers, followed by African-American men (more than 12), and U.S.-born Latino men (nearly 12), white men (11.8), and Asian Americans (nearly 10). Other ethnicities have a rate of around 11 per 1,000 workers. Historically, ethnic minorities have worked some of the most dangerous, dirty jobs and have been exposed to high levels of cancer-causing emissions. Often without adequate safety training and equipment.

The ‘haves’ and ‘have nots’

DEI issues and policies, and AI applications, are primarily the domain of multinationals that have the deep-pocket resources for AI uses in safety and health, the thousands of employees needed for viable data and patterns analysis, and the potential reputational risks warranting senior leadership buy-in of the need for workplace equity. 

But the U.S. economy is built on 33.3 million small businesses — 99.9% of all U.S. businesses, according to the U.S. Small Business Administration. In 2020, the majority of construction enterprises in the United States had less than ten employees, according to Statista.com. As of 2020, there were 473,108 small and medium sized specialty trade contractor businesses. These businesses consist of foundation, structure, building exterior, structural steel and precast concrete, poured concrete foundation, framing, masonry, siding, glass and glazing, roofing, building equipment, electrical and wiring, plumbing, heating and air conditioning, finish carpentry, tiling, flooring, painting and wall covering, drywall and insulation contractors and other contractors. 

How long for AI to trickle down?

How many years or decades will it take for AI applications to trickle down to small contractors? These are businesses on the far side of the digital divide. The odds are as AI advances, the OSH benefits will not be equitably distributed. The scoping review states, “OSH professionals must consider how the development and use of AI tools impact equity. Addressing the equity impacts of AI requires a paradigm shift for OSH professionals.”

I agree. Professionals out in the field, with frontline, first-hand knowledge of their workers, are uniquely positioned to ensure equal distribution of the benefits of AI that will potentially reduce the higher injury and fatality rates for minority workers. 

The challenge is, has been, and will be that OSH professionals are overwhelming out in the fields of large corporations that can afford their salaries. Few pros are employed in small construction and contracting businesses — that’s the exception to the rule. Who will ensure AI and other tech advances benefit at-risk workers on the far, far side of the digital divide? 

It will take a village, pardon the expression, a network of national OSH professional societies; local chapters; public health and safety-related NGOs; federal, state and local government agencies; tech vendors; organized labor; and high schools, tech schools and community colleges to reach the millions and millions of small businesses without safety and health knowledge or resources. This happens only if there is the will to make it happen — to make safety and health the true moral, ethical, “high ground” imperative everyone has always said it is.