How will AI change the work you and your employees do? The Manufacturing Leadership Council—the digital transformation arm of the NAM—is helping manufacturing leaders figure out the opportunities created by new generative AI technologies, including ChatGPT.
Recently, the MLC held a Decision Compass discussion to help manufacturers learn how to take advantage of these new tools safely and effectively.
The participants: The conversation was led by two members of West Monroe’s Center of Excellence for AI: Ryan Elmore and David McGraw. Elmore and McGraw shared their expertise and addressed questions from manufacturers throughout the call.
The use cases: AI is a diverse and complex tool that is likely to have a lasting impact on manufacturers across the United States. According to McGraw and Elmore, there are a range of applications for the technology, from supply chain optimization and production planning to predictive maintenance issues.
The workforce impact: According to Elmore, AI will also transform the manufacturing workforce.
- Some roles that involve repetitive tasks like data processing could be adjusted or eliminated, while some new jobs will be created around tasks like prompt engineering, which ensures AI programs deliver the most useful and accurate results. Most importantly, however, existing jobs will likely be modified to account for new tools.
- “Some are going to go away, some are going to be created, but the vast majority is going to change mentality, change infrastructure, change the way we work,” said Elmore.
Prompting success: Elmore and McGraw emphasized that the key to using generative AI effectively is developing useful prompts. How you ask AI programs for information, and what data you provide, will determine the quality of the output. They provided a few broad guidelines:
- Keep it simple: Your prompts should be detailed, precise and as succinct as possible.
- Data matters: The better and more detailed your data, the better your output will be.
- Keep it human: Generative AI still requires a human to determine the reliability of the output. Manufacturers shouldn’t plan to use outputs blindly without keeping a human in the loop.
- Share safely: Assume anything you put into AI that is not behind a paywall is not private. Only use data that you’re comfortable with others viewing.
- Follow up: If you receive outputs that don’t make sense, or that indicate some sort of failure, ask the program for more context and problem solving to assess whether the output is accurate or beneficial.
Safety first: AI can also be used in negative ways—for example, by cyber attackers attempting to gain private information from you using software that mimics the voice of someone you know.
- Elmore and McGraw emphasized that manufacturers using AI should consider providing trainings so employees can recognize and guard against safety issues.
The last word: “I think most importantly, you’re only limited to your imagination,” said McGraw. “There’s really a lot of use cases that can be solved with this technology.”
Learn more: Want to find out more about how digital tools are changing manufacturing? The MLC will delve deeper into these issues at this year’s Rethink Summit, taking place June 26–28 in Marco Island, Florida. Learn more and register here.