Revolution or Evolution? Exploring the Latest Breakthroughs in Industrial AI

ChatGPT and other Large Language Models have put AI in the mainstream consciousness with extravagant claims of a new era of AI, or even a 5th Industrial Revolution! The hype isn’t likely to stop any time soon with the Trillion Dollar Hyperscalers being joined by Nvidia in high profile 'AI wars.' 

With questions flooding into ARC from our clients, I recently embarked on a new adventure into the perplexing world of the breakthroughs in Industrial AI. Although armed with many years of expertise in IT, OT, and ET, I'm no expert in Artificial Intelligence. I plan to be your enthusiastic guide on my voyage of discovery, asking more questions than providing answers in this first installment of our blog series. Together, we'll navigate the twists and turns of AI's latest breakthroughs, making stops where we uncover repeatable business value. So, fasten your seatbelts and join me on this thrilling expedition, where we'll prove that even industrial engineers can learn new tricks in the ever-evolving realm of AI!

As I dove headfirst into this captivating underworld, I found myself tumbling down the rabbit hole of machine learning history, where I rediscovered neural networks (I used them in advanced process control solutions over 2 decades ago), the fascinating art of data annotation and labeling, and the myriad flavors of learning: unsupervised, semi-supervised, supervised, and even reinforcement learning from human feedback (RLHF). All this exploration led me to the doorstep of the much-hyped large language models (LLMs) and foundation models (FMs) that have taken the AI world by storm – think OpenAI's ChatGPT, Google's PaLM, and Amazon's Titan FMs, to name a few. These behemoths of the artificial intelligence realm have piqued my curiosity and left me eager to unravel their mysteries while sharing my findings with you. So, let's continue our journey through this enthralling landscape as we attempt to separate fact from fiction in the ever-expanding universe of AI!

Fear not, fellow industrial engineers! This expedition is tailored specifically for us, the hardcore problem-solvers striving to make a difference in a world of scarce resources. Our mission is to zero in on high-value business outcomes as we navigate the complex processes of designing, sourcing, manufacturing, selling, delivering, and servicing our sustainable products and services. We're not merely dipping our toes into the vast ocean of AI's broad societal impact; instead, we're anchoring our focus on the tangible applications and innovations that can revolutionize our industry. So, join me as we roll up our sleeves, put on our thinking caps, and dive deep into the fascinating world of AI-powered solutions, all the while remaining true to our roots and core values as dedicated industrial engineers.

In the upcoming chapters of this AI odyssey, I plan to tackle a myriad of thought-provoking questions that have been swirling around in my head. Together, we'll ponder:

• What data governance and 'data fabric' do we need as the foundations for Industrial AI use cases?
• Along with the new levels of digitization of industrial data, how do we need to manage the potential leakage of intellectual property (IP) and increased cybersecurity risks, by governing the use of AI in industrial organizations?
• Is Generative AI the most efficient tool for every industrial AI use case?
• What high-value industrial AI use cases will benefit most from Generative AI?
• Will Generative AI shift the balance of people, processes, and technology in industrial organizations?
• Can industrial AI systems bridge the skills gap in real-world manufacturing?
• How do we decide where we still need humans in the loop (HIL), and what industrial use cases could be autonomous?
• Is the carbon cost of generative AI ecosystems a step backward in our pursuit of more sustainable products and services?
• How must IT, OT, and ET convergence evolve to better capture industrial AI business value?
• Who will be the industrial software leaders who will bring their domain expertise to augment and outshine generative AI in the quest for true industrial intelligence?
• Are industrial cloud and industrial metaverse investments being sidelined by hyperscalars amid the race to harness AI's full potential?
• How is AI reshaping the industrial ecosystem, and who should we partner with?
• As we discover the limitations of Generative AI, and demands already surfacing in the EU and US for traceability, explainability, and ethical practices, what's next for AI? Causal AI? Explainable AI?

As we embark on this captivating journey together, I and the rest of the ARC team eagerly await your invaluable contributions. We invite you to share your questions, experiences, learnings, and solutions with us, as we believe that collaboration is the key to unlocking AI's true potential in the industrial realm. Your insights will not only enrich our collective knowledge but also help us navigate the ever-evolving world of Industrial AI more effectively. So, let's join forces and explore this fascinating landscape as a united community of passionate engineers, eager to learn, grow, and innovate together.

Follow Colin on Linkedin or reach out to Colin directly at cmasson@arcweb.com for more details on his Industrial AI research agenda.