TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A MIT Technology Review Business Lab report published October 8, 2026, examines how industrial companies can adopt more capable AI while managing safety, security and reliability risks. AVEVA chief technologist Arti Garg advocates keeping people in critical decision loops and establishing guardrails as AI extends into robots, drones and industrial software.
MIT Technology Review reported October 8 on how industrial operators can manage safety risks as foundation models, agentic AI and physical AI enable more autonomous work in plants, power systems and other high-consequence settings. In the interview, Arti Garg, AVEVA’s chief technologist, argued that deployment should retain human oversight and use defined guardrails, because AI decisions can affect physical equipment and critical infrastructure.
Industrial AI has long been used for specialized tasks such as predictive analytics, including identifying patterns that may precede equipment failures. Garg said AVEVA has worked on industrial AI for more than 20 years. The newer development, she said, is the spread of general-purpose foundation models, alongside systems designed to act through software or interact with the physical world. These tools could support more complex tasks than earlier, narrower applications.
The report describes data integration as one practical area of use. Industrial information may sit in separate sources, including equipment telemetry, service logs and engineering documents. Newer technologies can connect and correlate those sources to assist operators as they diagnose problems. AI-enabled robots could also gather information in hazardous environments, potentially reducing the need for workers to enter them. These are potential applications discussed in the report, not evidence that every facility has deployed them or that they have produced measured safety gains.
Garg said AVEVA’s responsible-AI framework emphasizes security, efficiency, human safety and oversight. Her position is that AI should support people rather than take over critical decision-making: organizations need to define where automated systems may act and where a human supervisor remains accountable. The interview also addressed AI’s environmental costs. Garg is involved in an IEEE working group developing a methodology to measure impacts across electricity, energy, resources, water and carbon; the report does not say that the methodology is complete or adopted as a standard.
Human Oversight in Physical Operations
The stakes differ from those of many digital-only AI applications. When software influences industrial equipment or directs a robot in a plant, a mistaken or poorly understood output may affect worker safety, operational reliability and infrastructure. The report’s central concern is not simply whether the systems can perform tasks, but whether operators can set boundaries, monitor decisions and intervene when conditions warrant it.
That makes governance part of the operating model, rather than a check added after installation. Companies considering autonomous tools may need to specify which decisions can be automated, which require review, and how responsibility is assigned. Those questions matter to workers and facility operators as well as technology vendors, particularly in environments where errors can have consequences beyond a failed software process.
Environmental accounting is another part of the deployment question. AI may help manage power systems as renewable generation grows, but the systems themselves consume electricity and other resources. A measurement method covering energy, water, materials and carbon could give organizations a more consistent basis for evaluating that footprint. The IEEE work described in the report remains under development, so its eventual scope and use are not yet established.
industrial AI safety monitoring systems
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
From Predictive Tools to Autonomy
Industrial AI is not a new field, but the types of systems being considered are changing. Earlier applications often addressed bounded tasks, such as analyzing operational data to flag possible equipment problems. Garg told MIT Technology Review that the recent arrival of general-purpose foundation models, physical AI and agentic AI has widened the set of tasks that could be automated or assisted.
The report cites Garg’s reference to a study suggesting industrial AI adoption rose by almost 78% over the prior two years. That figure is attributed to a study mentioned in the interview; the source material does not identify the study, its sample, the precise measure of adoption or its comparison methodology. It should not be read as a universal measure for all industrial companies.
The article is a Business Lab interview produced in partnership with AVEVA, a company with a commercial interest in industrial technology. Garg’s remarks therefore describe her perspective and the company’s approach; they are not an independent evaluation of AVEVA’s framework or of the performance of industrial AI systems across the sector.
“How do we leverage these technologies while maintaining safety, while maintaining reliable operations, while still being able to deliver on the promises of the new capabilities?”
— Arti Garg, chief technologist at AVEVA
As an affiliate, we earn on qualifying purchases.
Deployment Results Still Unspecified
The report does not provide evidence that the proposed safeguards have been independently tested across industrial sites, nor does it quantify whether autonomous robots or AI-assisted systems have reduced injuries, downtime or operating costs. Potential benefits remain forecasts in the material provided, rather than verified sector-wide outcomes.
It is also not clear how AVEVA’s guardrails are implemented in particular products or facilities, who audits them, or how responsibility is divided when an AI-assisted decision contributes to an incident. The adoption statistic cited by Garg lacks enough methodological detail in the report to assess what it measures. The status, timetable and eventual requirements of the IEEE environmental-measurement work are likewise unspecified.
As an affiliate, we earn on qualifying purchases.
Governance Before Wider Deployment
The next steps described are organizational as much as technical: companies will need to reconsider workflows, establish safeguards and decide how experienced workers remain involved as systems take on more tasks. Human oversight and clear operating boundaries are central to Garg’s proposed approach, although the report does not set out a common implementation schedule or binding industry rules.
The IEEE working group’s work on measuring AI’s environmental impact is another development to watch. The source does not give a publication date for the methodology. Meanwhile, the practical test for industrial AI will be how operators evaluate performance and safety in specific settings, including whether systems can be monitored and stopped when circumstances change. No particular rollout, standard release or deployment milestone is confirmed in the report.
predictive maintenance sensors for factories
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the development reported?
MIT Technology Review published an interview about safeguards for increasingly autonomous industrial AI, featuring AVEVA chief technologist Arti Garg. It is a report on deployment challenges and proposals, not an announcement of a new product or completed standard.
Why can industrial AI carry higher safety stakes?
Industrial AI can affect physical equipment and operations, rather than only producing outputs in a digital environment. An unexpected decision may have consequences for worker safety, reliability or infrastructure, according to the report.
What safeguards does Garg advocate?
Garg says organizations should emphasize security, efficiency, safety and human oversight, with rules defining when automated systems may act and when a person remains responsible for a decision.
Has industrial AI been shown to make operations safer?
The report discusses possible benefits, including robots gathering information in hazardous locations, but provides no measured results showing sector-wide reductions in injuries or other safety outcomes.
What is the IEEE working group measuring?
The group is developing a proposed methodology to assess AI’s environmental impact across electricity, energy, resources, water and carbon. The report does not state that the methodology has been completed or adopted.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
