Researchers develop framework to blend AI and human intelligence for process safety

Artificial Intelligence (AI) has grown rapidly over the past few years, and with it, industries are able to automate and streamline operations.

The feature article is AIChE Journal Identify the challenges and benefits of using Intelligence Augmentation (IA) in process safety systems.

Contributing to this research were Dr. Faisal Khan, professor and chair of the Department of Chemical Engineering at Texas A&M University, Dr. Stratos Pistikopoulos, professor and director of the Energy Institute, and Dr. Rajeevan Arunthavanathan, Dr. Tanjin Amin, and Dr. Zaman Sajid of the Mary Kay O'Connor Safety Center.

Additionally, Dr. Yuhe Tian from West Virginia University offered a new perspective on using AI in process plants from a safety perspective.

Khan said the basis of the research is using AI approaches to handle safety alongside humans, rather than replacing them in operational decision-making.

“This research aims to develop a comprehensive framework based on IA to integrate AI and human intelligence (HI) into process safety systems and ensure improved safety and efficiency,” said Arunthavanathan. “We aim to clearly understand the potential and limitations of AI and propose IA strategies for its effective implementation to minimize risks and improve safety.”

Helping humans, not replacing them

Khan believes that combining AI with human intelligence can allay concerns that AI may eventually replace humans as it becomes better at performing tasks.

“The study explores the challenges of incorporating AI technologies into real-world industrial applications and how IA can improve process monitoring, fault detection and decision-making to enhance process safety,” Amin said.

Khan argues that AI will improve safety by analyzing real-time data, predicting maintenance needs and automatically detecting faults, but IA approaches that use human decision-making also hold promise for reducing accident rates, lowering operational costs and improving reliability.

“The application of AI in chemical engineering faces significant challenges and falls short in ensuring comprehensive process safety,” Sajid says. “To overcome these limitations, IA is being deployed that works in tandem with human expertise rather than replacing it.”

The study identifies several risks associated with the implementation of AI and IA in the process industries. AI risks include data quality issues, over-reliance on AI, lack of contextual understanding, model misinterpretation, and training and adaptation challenges. Meanwhile, IA-related risks include human error in feedback, inconsistencies in AI-HI decision-making, biased judgments, implementation complexity, and reliability issues.

“The researchers are particularly intrigued by the challenges of AI and are conceptualizing IA to augment human decision-making in process safety,” Tian said. “They are fascinated by how AI can provide accurate and rapid responses based on data analysis, while human intelligence can provide broader insights and considerations, including ethical and societal factors.”

Khan believes this research highlights the importance of developing reliable, trustworthy and secure AI systems for industrial applications.

“AI and human intelligence working together are seen as essential to improving process safety,” Khan said. “Continued exploration of this synergy to meet the evolving demands of industrial safety will continue to enhance AI capabilities while ensuring a robust risk management framework is in place,” Pistikopoulos added.