Pairing industrial AI autonomy and human control

THE NEWS

What happened

Industrial AI is evolving from passive observation to active 'living digital twins' that maintain factory health through autonomous adjustments. While agentic AI can now act independently on specific tasks, human oversight remains critical for defining operational boundaries and ensuring safety. Major manufacturers like Saudi Aramco and Nestlé are already deploying these systems to optimize energy use and predict failures, balancing automation with necessary human judgment.

CONTEXT

Why it matters

Industrial AI is shifting from simulation to action. 'Living digital twins' now actively maintain factory health through autonomous adjustments, acting like an organism with its own metabolism. While agentic AI handles routine optimization—reducing energy use by up to 16% in some cases—human operators still define the boundaries of safety and strategy. The industry is moving toward a hybrid model where AI provides reflexes for homeostasis, but humans remain the ultimate guardians of operational integrity.

AT A GLANCE

Key facts

• Living digital twins combine high-fidelity digital models with the ability to intervene in physical operations for real-time control.
• Agentic AI is enabling industrial systems to achieve 'industrial homeostasis' through constant, small adjustments that keep plants healthy.
• A 2026 survey indicates 66% of manufacturers are using or planning to use agentic AI, with 40% expecting factory autonomy by 2030.
• Experts predict AI will operate in a 'supervisory' mode similar to SCADA systems for the foreseeable future, requiring human approval for most actions.
• Nestlé reduced energy use by 16% at Nesquik plants using AI optimization for industrial dryers.
• The EU AI Act classifies safety components in critical infrastructure as high-risk, mandating human oversight.
• Future developments include physical AI integrating robotics and sensing to monitor pipelines and power lines without human entry.