Honeywell Technologies CEO says 50 years of automation have gotten too complex for the humans running it
Vimal Kapur says “physical AI” is the fix for hospitals, data centers and refineries where people can no longer keep up with the exceptions.
Honeywell Technologies CEO Vimal Kapur warns that 50 years of automation have become too complex for the humans operating them. Speaking at the Fortune 500 Innovation Forum in Detroit, Kapur noted that while facilities often run themselves, often 95% of the time, a person is needed to handle exceptions. He attributes the complexity to the addition of more software and physical infrastructure over the years.
Kapur pointed to recent incidents, such as the FAA's nationwide ground stop and Microsoft's 365 outage, as examples of the complexity that humans struggle to manage. He emphasized that physical AI, which encompasses everything from hospitals to refineries, adds a layer of learning systems on top of existing automation to make humans more productive.
However, the main obstacle is the lack of accessible data from the physical systems, which is often not available in the public domain. Honeywell sells automation and its AI layer to customers in various industries, and in October, they launched Forge Cognition, an AI platform aimed at autonomous mode for buildings. Kapur emphasized that physical AI is about increasing productivity, not replacing jobs.
C.H. Robinson, a freight broker, also addressed the issue, describing it as "structural variability" in supply chains. Their solution, "Lean AI," involves AI agents for specific tasks to increase productivity, as seen in a 65% increase in shipments per person per day in their North American Surface Transportation unit since the end of 2022.
The main challenge is the shortage of skilled workers, with manufacturers needing up to 3.8 million additional workers between 2024 and 2033 due to retirement and other factors.
Written by urgent.news from Fortune's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.