June 21, 2024 4 Challenges for Transforming Manufacturing Through AI

AI Engineering has the potential to transform manufacturing. Here’s how industry leaders can drive that process.

By Pramod P. Khargonekar, ERVA Co-Principal Investigator; Distinguished Professor of Electrical Engineering and Computer Science and Vice Chancellor for Research, University of California, Irvine

The power of artificial intelligence to transform industry is often taken as a given. However, there is a gap between expectations and implementation that many industry leaders are not addressing. One essential element in the AI transformation equation is the need to proactively bridge two crucial fields: AI and engineering.

The strategic convergence of AI and engineering, envisioned as AI Engineering, represents a generational opportunity to supercharge engineering to transform manufacturing. This was the premise of a recent report, AI Engineering: A Strategic Research Framework to Benefit Society, on which I was the lead author. The report was the result of a two-day visioning event in late 2023 involving researchers, industry leaders, policymakers, and other stakeholders convened by the Engineering Research Visioning Alliance (ERVA), an initiative funded by the U.S. National Science Foundation (NSF).

AI Engineering is a two-way opportunity: Engineering disciplines will bring their domain knowledge, techniques, tools, and culture to create new forms of explainable, trustworthy, and reliable AI-enabled cyber-physical systems. At the same time, increasingly capable AI tools can transform the fundamental disciplines of engineering. They will also transform the major engineering endeavors of design, manufacturing, and infrastructure. To enable this breakthrough and realize AI’s full potential, leaders in government, industry, and academia must work together.

Simply put, the scope for AI Engineering to transform manufacturing is immense. AI Engineering will:

• accelerate the way engineers design products;
• optimize manufacturing operations; and
• create value for service systems built on top of engineered products.

Each of these opportunities holds enormous potential for progress.

As AI becomes industrialized, the process will encompass data science and analytics, machine learning, cyber-physical systems, digital twins, and beyond. As more sensors and smart analytics software are integrated into networked industrial products and manufacturing systems, predictive technologies can further learn and autonomously optimize performance and productivity.

The ERVA report provides four specific focus areas for industry leaders to explore for their organizations. (More detailed descriptions can be found in our report.)

1. Transform manufacturing through industrial AI Engineering. The key obstacles to successfully developing and deploying AI for industrial applications include the lack of good quality data and systematic approaches in developing and validating AI. Adoption is also hindered by challenges in AI trustworthiness and lack of standards. Industrial AI Engineering research will advance a systematic discipline that develops, validates, and deploys a range of AI approaches for industrial applications with sustainable performance, scalability, and security. Combined with state-of-the-art sensing, communication, and big data analytics, a systematic industrial AI methodology will allow the integration of physical systems with computational models. It will lead to an integrated learning platform to address the “3 D issues” – data, discipline, and domain – to strengthen interdisciplinary capabilities in diversified AI Engineering systems.
2. Design safe, secure, reliable, and trustworthy AI systems. AI safety has three distinct but complementary dimensions: (1) assuring a deployed AI system is safe and reliable, (2) using an AI system to monitor and improve the safety and reliability of a (potentially non-AI) system/platform, and (3) maximizing safety and trust in collaborative human-AI systems. Engineers must develop and monitor these systems in manufacturing with a focus on better understanding complex systems, ensuring good data management, safeguarding data, and ensuring regulatory compliance. And those responsible for AI implementation must prioritize safety and trust.
3. Build and operate AI engineered systems with cradle-to-grave state awareness. Cradle-to-grave system state (CtGSS) awareness is a novel concept with the potential to revolutionize the way engineered systems are conceived, designed, manufactured, deployed, operated, and retired. CtGSS approaches can only be realized through multidisciplinary technology development requiring expertise from almost all engineering and data science disciplines. This highlights the importance of investing in engineers with AI expertise and training.
4. Overcome scaling challenges in engineering. As we envision the future of AI Engineering, major cross-cutting and high-impact opportunity lies in scaling algorithms to very large problem instances and data sets. This illustrates the two-way nature of AI Engineering: AI development depends on engineering expertise, and engineering feats in manufacturing can be accelerated and expanded using AI.

Addressing these four challenges will depend on industry leaders making the research investments necessary to build this emerging discipline. As our report makes clear, government agencies and higher education have critical roles to play in reimagining the engineering profession and upskilling the workforce. To that end, manufacturing leaders can accelerate the drive to adoption by offering experiential learning opportunities in industry and curriculum development in partnership with education institutions. Transforming manufacturing through AI must be a collaborative strategy to successfully overcome the challenges and transform manufacturing through AI Engineering.

Pramod Khargonekar is Distinguished Professor of Electrical Engineering and Computer Science and Vice Chancellor for Research at the University of California, Irvine, and Co-Principal Investigator at the Engineering Research Visioning Alliance (ERVA), an initiative funded by the U.S. National Science Foundation (NSF). He earned a bachelor’s degree in electrical engineering in 1977 from the Indian Institute of Technology, Bombay, India, and master’s degree in mathematics in 1980 and doctoral degree in electrical engineering in 1981 from the University of Florida. He also served briefly as deputy director of technology at ARPA-E, U.S. Department of Energy in 2012-13. He was appointed by the National Science Foundation to serve as assistant director for the Directorate of Engineering in March 2013, a position he held until June 2016. In this position, Khargonekar led the ENG directorate with an annual budget of more than $950 million. In addition, he served as a member of the NSF senior leadership and management team and participated in setting priorities and policies.

Subscribe to Industry Today