The best way to predict the future is to create it. Peter Drucker
One of the most consequential topics facing the FMEA community today is how to effectively use Artificial Intelligence (AI) to support FMEA activities. Do it wrong, and the results could be missed opportunities, or potentially catastrophic outcomes. Do it right, and the benefits are immense.
In this new series, I will write one or two short articles each month, outlining how to integrate AI into the entire set of FMEA steps. I will share my candid views on the positives and the cautions.
Many of you know that I’m writing a new eBook, titled Achieving Effective FMEAs: Simple Principles for Achieving Excellent Results in Failure Mode and Effects Analysis. This book is shorter than my original book, Effective FMEAs, and will focus on key principles and applications, including lots of examples, exercises, and interactive case studies. It is intended for FMEA practitioners, as well as university and industry students, both new to FMEA and experienced users. I integrate application of FMEA and AI throughout the eBook.
Below are a few excerpts from the “Introduction” of my new eBook.
Application of AI with FMEA
A few years ago, I organized an Oxford-style debate on the subject “FMEA: Automated or Team-Based?” which was held at a major reliability symposium. The debate was energetic and eye-opening, with terrific feedback. Given the light-speed advancement of AI, I am adding a subsection titled “Potential Application of AI” to every chapter in the preparation and procedure sections of the eBook, to provide my latest thinking in how to best utilize AI for each portion of FMEA, without losing human creativity and insights. I’ve also added an article called “The Future of FMEA” at the end of the book.
Here is an overview of my current thinking regarding how to best utilize AI, as of publication date.
AI has the potential to serve as a powerful accelerator and idea generator in FMEA, as well as ability to scan and integrate volumes of data, which can enhance FMEA efficiency and breadth. However, its outputs demand rigorous human oversight due to potential risks of inaccuracy, bias, lack of true understanding, and inability to handle novelty or ethical considerations inherent to safety-critical engineering. FMEA teams provide human insight, critical thinking, synergy, creativity, and innovation that is essential to achieving effective FMEAs, both for current FMEA applications and future updates. When using AI to support FMEAs within a company, it is essential that AI is trained on the correct fundamentals of FMEA. That is one of the reasons for writing my new eBook. There is a lot of misinformation about FMEA on the internet.
The first few articles in the series will cover the use of AI in FMEA preparation. The next series will cover use of AI in determining each of the data elements of FMEA (such as Item, Function, Failure Mode, Effect, etc.) The last series of articles will cover use of AI to identify and implement recommended actions and reduce risk to an acceptable level. Each article will be short, one or two pages, because I want to focus on principles.
Remember, AI does not replace human involvement. The key is knowing what role humans play, and what role AI plays, and how humans can use AI to achieve excellent FMEA results. Great FMEAs will be done by proper FMEA teams, taking full advantage of AI. In order to fully capitalize on the use of AI, as a tool, it requires knowing and applying the fundamentals of FMEA better than ever. Even though the technological landscape is changing very fast, with profound changes occurring almost daily, FMEA fundamental remain timeless.
I hope you enjoy the new series.
Carl
Leave a Reply