It’s not just the rapid advance of artificial intelligence technologies, but the varying types of AI deployed by automation tech suppliers that are creating some confusion in the manufacturing automation marketplace.
This confusion has appeared in a couple of recent conversations. For example, I’ve heard dismissals of AI technologies based on news about the hallucinations in generative AI technologies that can affect the output provided. While this is true, the hallucination comments I heard were made in discussions about AI data analytics, not generative AI — often referred to as genAI — which is the type of AI tech that can “hallucinate” in its responses.
Even with genAI, there’s a big difference between genAI that’s been trained on specific, targeted sets of data relevant to a specific manufacturing facility and its equipment and software systems and genAI that’s been fed data on a wide array of subjects from a variety of sources — many of which may not have been dependable in the first place.
To help provide some clarity on this, I want to focus on the application of AI for data analytics and genAI for manufacturing production operations and how they interact with industrial automation technologies.
The difference between data analytics AI and genAI
Let’s start with AI for data analytics. While this is a relatively recent addition to the automation technology arsenal, it’s been in use for several years now for applications ranging from production analyses to predictive maintenance. In very basic terms, AI for data analytics in a manufacturing setting essentially crunches the data fed to it from a company’s plant floor equipment and software systems — including front office software such as ERP — and applies algorithms to sort through it all to highlight trends and anomalies and provide insights into business possibilities based on the correlation of data gathered from these disparate systems.
Generative AI can create original content — ranging from text, images, video, audio or software code — in response to a user’s prompt or request. Because genAI can be fed so much data from so many different sources, we see issues like hallucinations and other problems with the technology that require its results to be thoroughly reviewed by humans before being put them into action. But again, that’s general use generative AI. In a more controlled context, where the data input into a genAI system is provided by trusted sources and is focused on the equipment and systems specific to a particular company or group of companies working together, the results will be far more reliable.
That’s why you’re seeing so many automation technology companies implementing genAI technology to develop what’s commonly referred to as copilots. These systems are trained on a relatively closed set of data specific to the user’s application and the technologies they’re tied to rather than by scraping resources from all over the internet.
How genAI is being implemented by automation tech suppliers
Just as data analytics AI has become ubiquitous in manufacturing systems of all types over the past several years, we’re seeing a similar upswing now with genAI for manufacturing operations and design applications. A few recent examples include the announcement from Siemens and ServiceNow around their partnership to boost industrial cybersecurity and drive the integration of generative AI into shopfloor operations. This partnership involves the connection of Siemens' Sinec Security Guard for industrial vulnerability management and Siemens Industrial Copilot for generative AI-powered automation with ServiceNow’s workflow automation. Siemens genAI-powered Industrial Copilot will be used to provide users of ServiceNow’s Platform with new levels of control over operational processes through the interaction between static and dynamic machine data. The phrase “new level of control” refers to the ability of users to interact with the copilot technology in their own language to get detailed instructions and recommendations based on their requests. ServiceNow said that its ability to automate workflows — from maintenance scheduling to real-time problem-solving — helps ensure that the AI-driven insights from the copilot translate into tangible and efficient actions that improve productivity and minimize downtime.
Read more about this partnership between Siemens and ServiceNow.
And on the manufacturing design front, Tony Carrara with Rockwell Automation, recently wrote an article for Automation World about the use of genAI for generative design applications and how it’s impacting automation technologies used across the manufacturing industries. Tony said that generative design, which has long been used by automation manufacturers in the design of their products, is experiencing a significant evolution with the integration of genAI. He said genAI brings a new dimension to generative design by incorporating human-in-the-loop capabilities to transform how engineers and manufacturers conceptualize, create and optimize automation technologies.
He added that it's important to differentiate between existing generative design capabilities that use traditional AI and the emerging trend of integrating genAI.
Unlike traditional generative design methods that rely on AI algorithms alone, the addition of genAI introduces a more interactive and iterative approach, where engineers can provide feedback and guide the AI system towards more optimal solutions. Tony said this allows them to explore vast design spaces and generate numerous potential designs based on specified parameters, constraints and performance goals. He noted that this approach is particularly well-suited to automation systems, where multiple variables and competing objectives often need to be balanced.
Applying genAI-powered generative design to automation systems increases the speed at which multiple design alternatives can be generated and evaluated. Tony said that, in a matter of hours or days, the system can produce hundreds or even thousands of design options, each optimized for the given parameters. For example, if sustainability is a priority, the genAI can be directed to focus on eco-friendly design options.
Another application cited by Tony involves technology alignment with industry standards and best practices. He said genAI can be used to verify system compliance with cybersecurity standards by highlighting areas where a system deviates from established norms, helping engineers maintain consistency and quality across projects. The technology is also being used to standardize practices across engineering teams, particularly in scenarios where engineers of varying experience levels need to adhere to the same design standards and use consistent libraries. Tony added that this consistency is valuable when replicating systems across different sites or environments, as genAI can suggest appropriate adjustments while maintaining overall design integrity.
Keep an open mind about industrial genAI applications
The bottom line here is to not be too dismissive of new AI applications coming to automation technologies because of the issues we’ve all heard about with the genAI tools that get most of the media attention for general use like ChatGPT or Gemini. The genAI tools from automation suppliers focus on specific sets and sources of data to ensure the accuracy of results.
As a case in point for keeping your mind open about industrial genAI, consider that about 20 years ago many manufacturing engineers did not consider Ethernet to be a capable plant floor networking option.
Further development of genAI tech will be important to the manufacturing industry’s focus on capturing the knowledge of its expert engineering, operations and maintenance personnel to guide the next generation of industry workers. These manufacturing-focused genAI tools are poised to be the technology that make that goal a more easily attainable reality.
