Much of the focus on artificial intelligence now centers on its actual and potential capability to improve business productivity and profitability, and its threat to automate countless employees out of their current jobs. But a new analysis of workplace safety data broadens the scope of AI's likely impact, estimating that the new technology could lower occupational injuries by an average 6 percent annually over the next five years.
That forecast was the top finding from a recent study by Arizona injury law firm Lambert Goodnow. It crunched years of data from the U.S. Bureau of Labor Statistics, World Economic Forum, and other organizations to make two important conclusions. The first was that nearly one-third of all current work tasks are expected to be automated using AI by 2030, at an average rate of 22 percent across all economic sectors. The second was that as a result of that tech-driven change, accidents in U.S. workplaces are anticipated to decrease by nearly 6 percent each year over the same period.
"The predicted automation of 30 percent of tasks by 2030 is expected to reduce U.S. workplace injuries by 5.9 percent, preventing approximately 161,000 injuries annually within five years," the Lambert Goodnow analysis said, noting improvements will vary considerably across different jobs and business activities. "(W)hen looking at an industry-level breakdown, some of the most dangerous are likely to only become 2 (percent) safer."
The study examined workplace injury rates in various sectors and industries in both public and private businesses -- a blend that makes some comparisons challenging. But its overall conclusions show the AI safety improvement trend will likely affect a much larger number of employees than the forecast's percentages of declining incidence might suggest.
For example, injury rates at private healthcare companies are expected to drop by 6.3 percent through AI automation. While that decrease is only slightly higher than the expected 5.9 percent U.S. average, "this change could prevent nearly 30,000 injuries annually at the national level," the report said.
The study noted that forecasted the rate at which tasks are automated with AI over the next five years vary significantly across business sectors. They ranged from as much as 40 percent by administrative, support, waste management, remediation, professional, scientific, and technical services, to as low as 22 percent in arts, entertainment, recreation, accommodation, and food businesses.
But using a historically substantiated calculation that a 10 percent increase in automation has typically produced a 2 percent drop workplace injuries, the report said safety gains from AI would be considerable across industries and individual businesses, regardless of their adoption rates.
Still, the study indicated the biggest beneficiaries of AI workplace safety improvements are those likely to integrate it fastest over the next five years. But even sectors that are slower to embrace the new technology are expected to see injury rates drop. Those include agriculture, forestry, and fishing businesses; real estate, rental, and leasing companies; and finance, management, and insurance firms.
Similarly, the report said sectors with lower potential for introducing AI tech are still expected make significant workplace safety gains in simple human terms.
"Arts, entertainment, and recreation, for example, is predicted to see a 4.3 percent drop in injuries," the study said, noting that would be 1.5 percent lower than the national average. "A 1,600 drop in injuries annually in the industry is, however, still an impressive figure."
Despite the forecast of increased AI-linked safety improvements over the next five years, the report said the tech won't eliminate the risks in businesses or sectors whose activities lead to higher injury numbers in the first place. It also won't alter their individual incidence rates to the same degree.
"Four of the ten most dangerous (professions) have low automation potential and are likely to remain at least 97 percent as dangerous as they are today." the study said of businesses whose large workforces will limit how much their injury per 100 employees rates decrease. "For example, while the national average injury incidence rate is projected to fall to 2.29, the rate in state-run nursing and residential care will only fall to 8.7, which is close to four times higher. Others, like couriers and scientific professional services, are far more automatable, but also are significantly larger than some others on this list."
Still, even small declines in accidents from AI automation translate into thousands of employees being spared injuries and deaths that would have happened otherwise.
