AI Robots Run 24/7 But Still Need Human Oversight, Study Finds

The limits of autonomous manufacturing

The vision of fully automated factories running without any human presence remains elusive, despite significant advances in robotics and artificial intelligence. While AI-powered robots can operate around the clock in highly controlled environments, they fundamentally cannot function entirely without human oversight.

The core limitation centers on what researchers call tactile intelligence—the ability to assess physical situations, calculate risk, and adapt to unexpected conditions. Robots currently lack the sensory feedback systems needed to handle soft materials or make the nuanced judgment calls that human workers perform instinctively.

Why it matters

As companies invest billions in automation technology, understanding these constraints helps set realistic expectations for return on investment. The 90-10 rule emerging in modern manufacturing—where robots handle routine tasks while humans manage exceptions—represents a more practical path forward than pursuing completely unmanned operations.

Where robots excel

Autonomous systems perform exceptionally well in environments engineered for predictability. According to Massachusetts Institute of Technology experts, massive robotic fleets successfully move shipping containers 24/7 in ports and industrial warehouses because operational parameters remain strictly controlled.

These so-called "lights-out" factories demonstrate that robots can manufacture products continuously when conditions are structured and repetitive. However, even these facilities require human engineers for software maintenance, system upgrades, and handling unexpected situations.

The autonomation approach

Rather than eliminating humans entirely, modern industry has adopted a concept called autonomation. This framework assigns AI robots to handle 80 to 90 percent of mundane physical work, while skilled human operators manage the remaining tasks requiring complex decision-making.

This division of labor acknowledges a fundamental gap in current robotics: despite remarkable progress in visual perception, machines struggle to adapt when confronted with unfamiliar physical situations or system failures.

Teleoperation as backup

When automated processes break down, robots lack the improvisational capacity to recover independently. Industries address this limitation through teleoperation, where human workers remotely control machines to provide the quick thinking and risk assessment that artificial intelligence cannot yet replicate.

This hybrid model—combining robotic endurance and precision with human adaptability and judgment—appears to represent the realistic state of automation for the foreseeable future rather than a temporary compromise.

These findings were first reported by Automation Watch.