Autonomous Operations Will Require Causal AI

Overview

Interest in autonomous operations continues to grow spurred by the recent pandemic, but also dynamics in the workforce and a drive to improve ESG (environmental, social, and corporate governance) rankings. Manufacturers have begun layering artificial intelligence (AI) on top of advanced process control and regulatory control to increase the autonomy of their processes. Unfortunately, the autonomous plant is still elusive.

One obstacle encountered is the inability for trained AI models to adapt to conditions they have not been trained with. Causal AI is an artificial intelligence system that can explain cause and effect. This technology is used by organizations to help explain decision making and the causes for a decision. Systems based on causal AI, by identifying the underlying web of causality for a behavior or event, provide insights that solely predictive AI models might fail to extract from historical data. An analysis of causality may be used to supplement human decisions in situations where understanding the causes behind an outcome is necessary, such as quantifying the impact of different interventions, policy decisions or performing scenario planning.

Employing causal AI holds the potential to enable more comprehensive AI models to adapt to new conditions and increase autonomy in manufacturing.

Greater Focus on Autonomous Operations

Interest in autonomous operations continues. Though the memory of the need made apparent by the pandemic fades, the benefits continue to resonate. Fortunately, companies found they have much of the necessary automation foundations in place. These include the sensors, control systems, advanced process control, historians, and manufacturing execution systems. Prior to the pandemic, these systems were generally implemented to meet a set of operating conditions that relied more heavily on the immediate availability of personnel. During the pandemic, technologies that enabled remote access and better automation went from luxury items to necessities. Additionally, these technologies revealed their shortcomings and received some much-needed investment for improvement.

End users looked at easy wins by applying the inherent automation capabilities of the installed control systems. Then the work in this area tended to push greater dependence and focus on artificial intelligence (AI). However, industry applications for AI continue to have limited scope as end users learn the limitations of the technology and themselves. As many learned, correlation does not mean causation.

The Challenges Manufacturing Faces

One of the biggest obstacles to the use of AI for autonomous operations is error recovery or resilience. Most AI cannot reason beyond what it’s been trained on, whereas a human can reason beyond their experience because humans tend to understand the ‘why.’ Should operating conditions deviate from what the AI has been trained to expect, a human must intercede.

This is effectively the largest advantage a well-trained human operator has, but this operator experience and reasoning is encapsulated in the brain of the human. This experience aspect is also why when certain deviations occur, the response could vary based on the expertise of the operator(s) on shift.

Unfortunately, manufacturers are experiencing challenges with human staffing in general. Recently, ARC published an article on this topic depicting the challenges faced and some solutions available to address them including knowledge capture and autonomous technologies.

ARC Advisory Group clients can view the complete report at  ARC Client Portal

If you would like to buy this report or obtain information about how to become a client, please  Contact Us

Keywords: Autonomous Operations, Autonomous Plants, Artificial Intelligence, AI, Causal, Knowledge Management, Knowledge Capture, Predictive Analytics, ARC Advisory Group.