From factories to operating rooms: how AI is teaching robots to work safely alongside humans

The next phase of robotics hinges on a fundamental challenge: how to let machines work independently around people without endangering them. Humanoid robots are already reshaping manufacturing and healthcare, but unlocking their full potential requires clearing both technical and safety hurdles.

Industrial robots stepping out of isolation#

Agility Robotics has kept its Digit robots cordoned off in separate workspaces to protect human workers. That boundary is about to blur. The company plans to debut its Digit v5 model within the next year as the industry's first AI-powered humanoid robot designed for safe, direct collaboration with people on the job.

Both Agility and Boston Dynamics are contributing expertise to shape international safety protocols. They've joined forces in a standards-setting group working on ISO 25785-1, a proposed framework for industrial mobile robots. The draft is now under review by a technical committee focused on robotics. If approved, it will advance to a vote among the ISO's 89 member nations.

Surgical robots: autonomy with guardrails#

The medical field offers a window into how autonomy can coexist with safety constraints. Surgical robots, despite their precision, remain far from fully independent. "Surgical robots don't perform surgery automatically with full autonomy," explained Bhushan Patel, principal technical program manager at Intuitive, a surgical robotics firm in Sunnyvale, California. Instead, he described them as "still overwhelmingly human-driven systems with varying levels of intelligent assistance."

Robotic surgery began in the 1980s with a goal to aid astronauts in space, according to a piece by Patel in IEEE Transmitter. The concept soon evolved into terrestrial applications, including robot-assisted removal of prostate tissue, where surgeons guide robot arms with superhuman precision through minimally invasive techniques.

The rise of AI-assisted surgical tools#

Early robotic surgeons, from the late 1980s through the 2000s, were entirely teleoperated—surgeons planned and executed every move. Autonomy began creeping in as systems adopted what experts call "level one" capabilities: automatic motion stabilization for steadier cuts, safety boundaries that protect patients and staff from harm, and AI-powered image analysis that helps surgeons distinguish between tissues and organs. These advances let humans remain in control while machines handle precision and safety enforcement.