Future Shock, Adaptability through GenAI, and Architecture in Manufacturing

Introduction

“The inability to speak with precision and certainty about the future is no excuse for silence…it is more important to be imaginative and insightful than to be one hundred percent ‘right”. – Alwin Toffler.

In his 1970 book “Future Shock”, Alwin Toffler presents many walks of life in which disruptions would take place when the industry was in the beginning stages of “Information era” or Industry 3.0. Aspects those he spoke about were widely received with surprises since many people couldn’t imagine the outlook that Toffler outlined. As we are already in Industry 4.0 era in which Cloud Computing, Data and AI play the center stage, many people wonder what would happen with their lives in general through the advancements brought about by these technologies. Yuval Noah Harari in his book “Sapiens” addresses this question of how human civilization evolved over millions of years through continuous adaptation and reached where we are today while it faced continuous threats that sowed opportunity for improvement. The world of today is giving the human society the necessary technological ammunition required to continuously seek and improve.

Adaptability and Resilience through GenAI

Gen AI offers its practitioners a creative content based on the data that it is trained upon, often provides this content through prompting, like Google Search. It can be in the form of text, video, picture, and other forms of content creation and deliver.

Based on what is stated by Alwin Toffler himself, it is important to be creative while being moderate and need not be 100% right at the first try and instead the improvement should be iterative. The process of continuous learning helps the overall eco-system to evolve on need basis while addressing day-to-day requirements. If a company does work on such tiny process improvements, the result would be immense.

Some of the GenAI use cases those can bring value are Digital Twin and Process Improvement, Product Development, Intrusion prevention, Human Safety, and improvements in Key-Performance-Indicators.

Digital Twin and Process Improvement

Digital Twin replicates the exact production scenario through IIOT enabled devices and other supporting devices and platforms. It allows the employees to “play around” with it in a way that all production aspects could be visible while it doesn’t impact the actual process. Having a GenAI model embedded into it allows the Digital Twin to be more creative than what it can offer with while enabled by traditional AI models. This helps the production teams to understand various scenarios those couldn’t be thought of with traditional / ML based simulation approaches. This is applicable in both SKU based industry and Mass flow industry.

Product Development

Simulation, which is the core of product development to consider the conditional parameters while giving the Targeted output can be augmented with GenAI models. As seen above with Digital Twin, the scenarios that one can thought of can be exceeded quite fast to provide good business benefits. This also allows both OEM and its supplier eco-system to share their data on trust basis while also benefit each other on the amalgamated models. This can give business benefits for both the companies.

Human Safety

As it is recently demonstrated by OpenAI through its product “Sora”, a complete production scenario at different times and seasonality can be used as feed to GenAI system to get scenarios in which a potential danger can occur while Humans are involved in a particular process. This can help avoiding industrial events those can lead to potential loss of production or health issues. As demonstrated by ALCOA through their implementation of “Keystone Habit” in their production processes, a significant improvement can be brought into the profitability by addressing Health & Safety dimension. H&S not only improves the workplace safety but also improves overall environment that can eventually improve the process. Hence, by bringing controlled creativity within the shop floor can help improving the overall company performance and GenAI can play a major role.

Improvement in KPI

Improvements that an executive immediately look for is in the KPIs of his / her company that is reported to the Board. When the concept of Data Lineage is properly embedded into the KPI reports and the eco-system is powered by GenAI models, the executive can expect some good feedback on where and how the process improvement can be achieved. In case of traditional ML models, there will be many steps that needs to be taken care of before any such achievements can be made.

Architecture in Manufacturing

High level architecture presented above is applicable in manufacturing sector. Sensors those are coming from various plant areas can stream data on continuous basis to the LLM eco-system. Retraining of the models can be scheduled based on the learning of the designers of this system. Graphical Processing Units (GPU) are required to train and serve LLM models. Though the initial investment would be high, they can bring in substantial value over time. The output LLM model / finetuned LLM can be the base of a Digital Product that can address a use case. Typically, the user shall leverage such a product through a prompter like how internet users seek ChatGPT / similar tools to get their answers.

Conclusion

While the development of a GenAI system and its business benefits are presented in an abstract fashion, it is a vast field that evolves day-by-day. Its intricacies are complex at least for beginners. As Toffler stated, it is important to be imaginative and continuously better to adapt for the benefit of the society at large.