Honeywell to miners: build AI on existing plants

The summit, organized by AmCham Chile and Honeywell with sponsorship from Microsoft, framed mining's AI push against a straightforward demand story: data centers need electrical infrastructure and minerals such as copper, and mines want AI to lift production. Honeywell's José Simón tied the difficulty to geology — falling copper concentration in rock means more material must be processed for the same metal — and to a shortage of specialized personnel and the cost of unscheduled downtime. His remedy was not a new AI product but the expansion of existing tools: advanced process control, training simulators, remote operation centers, and equipment performance systems.

A second thread at the event concerned capability rather than equipment. Mining Council's Joaquín Villarino said the capacity to acquire technology far exceeds the capacity to use and leverage it, and pointed to the risk of knowledge loss as experienced personnel retire, requiring new profiles and preserved experience across both miners and suppliers. Honeywell's Jason Urso offered "digital cognition" — sensors, process models, and AI reading maintenance logs, alarm sequences, and operator actions — as a way to retrieve past responses for less experienced workers, while cautioning that AI can err.

Worker safety and resource limits closed the discussion, with Honeywell's Armando Pazos describing local measurement and centralized monitoring for gas exposure, and José Magalhães Fernandes noting that producing more requires more water, energy, and workers, and that technology must help close the gap between production and demand. Whether that gap narrows may depend less on what mines buy than on whether their boards treat the change as a management task.