Amazon explains robot rollout as researcher warns machines can dictate work pace

Amazon says warehouse robots spare employees the toughest physical work. But in interviews published by Ars Technica on October 12, researcher Brian Greenwood warned that machines working beside people can also dictate their pace. Amazon Robotics vice president Scott Dresser described a growing robotic-arm deployment, putting that tension alongside the company’s plans to automate more warehouse tasks.

Less lifting is not the whole safety question

Greenwood, a business professor at George Mason University, drew a distinction between older automation and collaborative robots. Historically, he said, automation often removed a dangerous task from a worker and performed it elsewhere. Machines that share the workplace change both the physical distance and the relationship between human effort and machine speed.

That concern sits within an already hazardous industry. The US warehouse sector’s injury and illness rate has typically exceeded twice the all-industry average, according to Occupational Safety and Health Administration figures cited by Ars. Amazon’s position is that robots reduce exposure to tasks carrying the greatest risks of fatigue and injury.

The scale makes the human-machine relationship consequential. Amazon operates more than 1 million warehouse robots, while its 2025 annual report counted more than 1.5 million full-time and part-time employees. Robots still cannot perform the full range of tasks people handle, leaving a large workforce operating alongside the machines rather than a fully automated warehouse.

The rise of collaborative robots really changes that dynamic for two reasons: one is that it is physically proximate, and two, the robot has the ability to dictate the pace of work.

Brian Greenwood, George Mason University business professor, speaking to Ars Technica

Robotic arms are a small but growing share

Dresser told Ars that Amazon’s deployed robotic arms number in the single-digit thousands. That is a much smaller category than the wheeled mobile robots that dominate its fleet. CFO Brian Olsavsky said on a July earnings call that Amazon expected to more than double its existing robotic-arm fleet during 2026.

The arms take on different pieces of fulfillment work. Robin sorts individual envelopes and boxes from cluttered piles, using suction cups to lift packages. Cardinal places packages weighing up to 50 pounds into rolling carts. Dresser said Amazon has just begun deploying Cardinal in larger numbers and will continue over the next several years.

Sparrow tackles a different problem: consolidating products from partly filled storage totes into densely packed ones. Dresser described it as Amazon’s first deployment of foundation models—AI models designed to support a broad range of tasks—in physical robotics. He said they help identify varied products and develop grasping strategies without manually coding every item.

Touch opens another set of warehouse tasks

Amazon’s newer Vulcan arm adds force-feedback sensors to vision, giving it a sense of touch as it retrieves or stores products in fabric-covered inventory pods. Dresser explained that inserting an item, making room for it or pulling it out is difficult for systems that rely on suction without touch feedback.

Vulcan began operating in Spokane, Washington, and Hamburg, Germany, in 2025. It is not yet widely deployed. In his deployment account, Dresser said Amazon expects to scale its use in 2027—a future rollout, not a completed milestone.

Amazon offers a concrete example of the physical relief it promises. Vulcan handles items in the top rows that otherwise require employees to use a step-ladder, as well as items just above the floor. The company says that lets people concentrate on picking and storing products within a more comfortable reach.