AI Governance Institute
← News

AVEVA's Industrial AI Framework Sets a Human Oversight Benchmark for OT Operators

What happened

On October 8, 2026, AVEVA chief technologist Arti Garg outlined the company's Building a safer path to autonomous industrial AI governance approach in MIT Technology Review. The article covers AI governance for industrial environments. The framework treats AI as a tool that augments human judgment rather than replaces it, and sets explicit rules about where automated systems may act on their own. Garg describes guardrails applied to foundation models, physical AI, and AI agents operating in mission-critical settings such as energy, water, and manufacturing plants. The framework also addresses security controls for environments where a wrong automated action can cause physical harm. AVEVA is also participating in an IEEE working group on AI environmental measurement. The group is developing a standard methodology for measuring AI's use of electricity, water, and carbon. This adds an environmental accountability layer to its governance approach.

Why it matters

  • ·Industrial operators using AI in mission-critical environments face growing scrutiny over whether human oversight is substantive or merely nominal. AVEVA's published framework, which names specific boundaries on autonomous action, raises the standard against which regulators and auditors will measure peer organizations.
  • ·The framework's explicit separation between AI-assisted decisions and AI-autonomous actions directly informs the design of human approval gates and escalation procedures. Compliance teams in operational technology (OT) sectors need to document where their own AI deployments sit on that spectrum. Regulators including those applying the EU AI Act (Regulation (EU) 2024/1689) are asking exactly that question.
  • ·Participation in an IEEE environmental measurement working group signals that AI energy and water consumption will become a reportable metric. Organizations without a process for tracking AI-related environmental costs risk being unprepared when disclosure requirements follow.

Stay ahead of stories like this

Get every Global AI governance development like this one, plus the rest of the week's developments. Every Thursday.

Free every Thursday. Unsubscribe anytime.

Governance controls affected

What to do now

  • ☐Map every AI deployment in your operational technology or industrial environment and classify each one as human-assisted or capable of acting without human approval at each decision step.
  • ☐Review your human approval gate documentation to confirm it specifies which actions require sign-off before the AI proceeds, not just after the fact.
  • ☐Ask your engineering team to confirm that AI agents in industrial settings cannot exceed their defined task scope without an explicit re-authorization step.
  • ☐Begin tracking AI system energy and water consumption data now, ahead of expected IEEE and regulatory reporting requirements, so your environmental disclosure program is not built from scratch under deadline.
  • ☐Request from each industrial AI vendor a written description of where their system can act autonomously and what stops it from exceeding those boundaries, then compare those descriptions against your own governance documentation.

What to watch next

The IEEE working group that AVEVA is contributing to has not yet published a draft standard for AI environmental measurement. Compliance teams should monitor its progress, as published metrics could quickly become a baseline for voluntary and then mandatory disclosure. Separately, regulators applying the EU AI Act (Regulation (EU) 2024/1689) are actively examining high-risk AI deployments in critical infrastructure sectors. Enforcement patterns from those inspections will clarify how much documentation industrial operators need to show that human oversight is real rather than procedural.

Related Coverage

Corporate Policy2026-11-16

IRS Pilot Shows Where AI Ends and Human Judgment Begins in Fraud Cases

The IRS is piloting an agentic AI tool to handle clerical steps in fraud case processing. This includes organizing taxpayer correspondence and summarizing responses. Human employees make all final decisions. The pilot is in development without a confirmed deployment timeline. The design explicitly reserves all case judgments for human staff.

Research2026-11-16

Regulated Industries Need Agentic AI Governance Infrastructure Now

DIA Global Forum's November 2026 article presents a five-stage capability maturity model for agentic AI in regulatory affairs. It draws on RegASK benchmark data. Most organizations take one to four weeks to respond to new regulatory requirements.

Research2026-11-12

OCC Supervisory Framework Puts Generative and Agentic Bank AI Under Examiner Scrutiny

The OCC issued a supervisory framework on November 4, 2026. It guides examiners reviewing how national banks and federal savings associations use AI in credit decisioning, payments, fraud detection, customer service, and risk management. The framework is non-binding and sets no new enforceable standards. It fills a governance gap left by the April 2026 interagency model risk management guidance, which explicitly excluded generative and agentic AI as 'novel and rapidly evolving.' Banks with material AI deployments in those categories now face structured examiner scrutiny without previously having been required to build matching controls.