Dive Brief:
- Drivers and warehouse-related jobs are most exposed to being taken over by robots, an analysis by AI company Anthropic found.
- The technology for autonomous cars has advanced enough that it’s possible for robots to do about 97% of the work of taxi drivers, making this job the most exposed. But it’s not just technological capabilities that determine if and when robots take over a job; it’s also whether robots can do it at less cost, adjusting for factors like reliability, error rates and speed, and whether people are ready to accept having robots do the work and if there are regulatory restraints.
- On a technological basis, robots are capable of doing three-quarters of the physical work that’s done in the United States today, making up 34% of working hours, but the use of robots is cost-competitive in only 0.3% of cases, the equivalent of about 300,000 jobs, the analysis found. As technologies improve and costs come down, robots could take over more work. “AI-powered robots could leapfrog our scale and do work they cannot today, for example by learning to climb ladders or use their arms and grippers more deftly,” the analysis said.
Dive Insight:
The analysis categorized robots based on the environment they’re capable of working in. Robots that work in structured environments, like an assembly line, can do the highest share of physical tasks — almost 50% — while those that work in an unstructured environment, like a construction site, can do only less than 2% of tasks. In the middle are facilities that people work in but that nevertheless have structures that can accommodate robots, like a warehouse, where robots can do about 22% of tasks.
Taxi drivers have an exposure rate of 2.2, the highest of all work, on a 0-3 index scale. “Autonomous vehicles have yet to upend driving jobs at scale, but their capabilities suggest that these jobs are at higher risk of automation than others,” the analysis said.
In addition to taxi drivers, at-risk driving-related jobs include those involving industrial trucks and tractors, paving and surfacing equipment, agricultural equipment, mining equipment like excavating machines, light trucks, shuttles and limousines.
Warehouse packers and packagers are other highly exposed jobs. There are robots that can do a range of work that packers and packagers do — pack containers, move materials around warehouses, visually inspect goods, erect cardboard boxes, print and apply labels and seal packages.
These jobs are at near-term risk because robots can do much of this work cost-effectively. The typical packer earns about $49,000 a year. A robot that’s capable of doing the work equivalent to 14 packers would cost less than the workers over the 10-year expected life of the robot after factoring in a $2 million purchase cost and $45,000 in annual operating costs, the analysis says.
The extent to which a job is at risk is nuanced, the analysis shows. Drywall specialists on a construction site, for example, face little exposure if they do the initial taping but a lot of exposure if they do the finishing work. The tape and “mud” that workers use to cover drywall seams would damage robot components, so that work is unlikely to be automated. But workers who finish the surface could face significant exposure. More than 40% of these “finisher” tasks — like spraying coats of sealant and sanding walls — can be done by robots, the analysis said.
Few of the jobs outside of environments specifically structured for robots are cost-competitive today, but that could change as costs for the technology and hardware come down. If costs dropped 20%, then the 300,000 jobs that are cost-competitive today would increase to 2.8 million. Based on historical data, it could take seven years to reach a 20% drop in costs. “Industry and government data suggest that robot prices have declined roughly 3% per year since the 1990s,” the analysis says.
Jobs with a physical component that are least at risk are those that take place in an unstructured environment and require dexterity that robots are unlikely to have any time soon. Although ditch digging would seem to be vulnerable to robots, the job in part requires a lot of dexterity. About 20% of the work requires “careful digging by hand around existing buried pipes, flowlines, cables and conduit,” the analysis said. “These activities require different sorts of physical abilities, like navigation and strength (linear trenches) compared to fine motor dexterity and perception (digging by hand around pipes).”
Physical jobs that require interpersonal skill are also at little risk. Teaching people how to dance, for example, while a physical job, requires cognitive and interpersonal skills. These skills are needed in about 54% of the job’s work time.
If the past is any guide, said the analysis, referencing historical cases of technology taking over work, “taxi drivers and warehouse packers will see changes sooner than nurses and mechanics. We expect that physical work will first be automated where robots have a foothold today.”
