“It’s all about speed” – Georgia-Pacific on artificial intelligence generation at the edge of industry

Like virtually every other large global company, paper products manufacturer and distributor Georgia-Pacific is working to leverage generative artificial intelligence (AI gen.) to increase workforce productivity. While the focus is “all on speed right now,” says vice president of innovation Michael Carroll, the underlying problem dates back to the 2000s, when manufacturing productivity growth began to slow and then plateaued as skilled workers began leaving the workforce in droves, which was an attempt to replace institutional memory with technology.

At the time, Carroll said, “You could start to find it difficult to stay productive… and that led to the advent of automation.” However, this led to overcorrection, where frontline workers were distracted by an overabundance of information and analysis that was intended to advance human knowledge. The cumulative effect of these distractions, this “generational vulnerability” to distractions, was counterproductive – instead of workers using technology instead of knowledge to become more productive, the opposite happened. Then the Covid-19 pandemic further exacerbated the problem of knowledge loss and generational vulnerability.

Carroll said that at its core, Gen AI is a technology that enables humans to interact with other technologies; it makes knowledge that exists in silos within the company something that can be collected and used to make decisions. It’s all about giving the employee contextual information when and where it’s needed.

He referred to the Purdue model, a reference enterprise architecture based on a six-level hierarchical model. Let me remind you:

  • Level 0 is a physical process that includes sensors, actuators, and other devices on the shop floor.
  • Level 1 refers to basic controls – such as programmable logic controllers and remote terminals that directly manage physical processes.
  • Level 2 is concerned with monitoring and supervising process control systems; is your first level of data aggregation and accompanying visualization.
  • Level 3 are manufacturing operations systems that support the overall operation of the manufacturing process, including scheduling, batch management, and quality control.
  • Level 4 deals with business planning and logistics, where manufacturing operations integrate with the manufacturer’s larger operations.
  • Level 5 is the enterprise network and covers everything from corporate IT and finance to human resources and company-wide applications.

Right now, Carroll said, people act as intermediaries between these levels and create a sort of “learning loop” that influences the decision-making process. “What we’re really trying to figure out now is, how do we build artificial intelligence that transforms the human learning loop… into an agent-based digital learning loop?” This future state will provide transparency between these layers in terms of understanding the fundamental inputs to the thinking and reasoning processes that result in greater intelligent decision-making outcomes. “Everything can be explained at the level of the generative AI model that your employees teach.”

Carroll continued: “In the agent world, you have to be able to convey to the employee what you need to know, what you need to do with it, and when you need to do it?… You have to be good at being able to deliver the content to the person in relation to what their personality says they are trying to achieve… The disappointing thing is that today you have to build this world yourself. There is no app that allows this. We hope that in the future there will be a platform where agents can live.”

Appreciating all the point solutions he sees vendors bringing to market, Carroll said: “The problem is organizing the problems and getting them out contextually and with priority on the front lines when they need to get there… You have to become your own context coordinator.”