Dr. Jan Liphardt highlights modular AI systems as key to safe, transparent robots entering human workplaces

Dr. Jan Liphardt, a researcher in physical AI systems, emphasized that robots built from modular AI components (smaller, specialized decision-making units rather than one monolithic system) perform more reliably and can be audited for safety — a critical shift as companies like Tesla and Boston Dynamics prepare to deploy humanoid robots in factories and warehouses alongside human workers.

The modularity approach lets engineers understand exactly why a robot made a specific decision (transparency) and catch safety failures before they happen, unlike black-box AI systems where even creators cannot explain the reasoning — this reduces liability and regulatory friction for companies deploying robots in regulated industries like healthcare and manufacturing.

For business leaders evaluating robot investments: modular AI architecture is becoming a non-negotiable requirement from insurers, regulators, and safety-conscious enterprises, meaning vendors without transparent, auditable systems will lose contracts to competitors who can prove their robots won't malfunction in unpredictable ways.