Section [3]: Integrating Artificial Intelligence with Human Decision-Making in Process Safety Risk Management: Strategies, Considerations➡️ ...
Published on by Hossein Ataei Far, Deputy Manager of the Research, Technology Development, and Industry Relations Center at NWWEC
➡️ Introduction
Explores the evolving role of Artificial Intelligence (AI) in process industries, emphasizing the importance of human involvement for safety. While AI offers significant potential in automating processes, predicting failures, and optimizing operations, its complete takeover of process safety is cautioned against. Instead, Intelligence Augmentation (IA), where AI and human intelligence (HI) collaborate, is recommended as the optimal approach for maintaining safety.
🔄Safety Layers: AI could enhance certain safety layers, such as Basic Process Control Systems (BPCS) and Safety Instrumented Systems (SIS), by automating routine tasks. However, human oversight remains crucial, particularly in layers requiring complex judgment, like emergency response and mitigation.
🔄IA Approaches: Three IA approaches are discussed:
🔹AI-in-the-Loop (AIiL): AI assists humans, who make final decisions.
🔹Human-in-the-Loop (HiL): Humans constantly monitor and influence AI, ensuring decisions align with broader considerations.
🔹Human-on-the-Loop (HoL): AI operates autonomously but seeks human input when facing uncertainty.
🔄Challenges: The implementation of fully automated AI systems faces challenges, such as data quality, model accuracy, and the need for human intervention during emergencies. A fully automated system without human involvement is deemed unsafe.
🔄Risks of AI-HI Collaboration: Potential conflicts between AI and HI, due to differences in decision-making processes, pose significant risks. These include AI misguidance from inaccurate data, human misinterpretation of AI decisions, and conflicts arising from differing intelligence types.
➡️ Conclusion: While AI holds promise for improving efficiency and safety in process industries, the keynote stresses the necessity of a balanced approach where AI supports human operators, ensuring safer and more reliable operations. IA, rather than fully automated AI, is presented as the safer path forward in process safety.
Reference:
[1] Arunthavanathan, R., Sajid, Z., Amin, M. T., Tian, Y., Khan, F., & Pistikopoulos, E. (2024). Process safety 4.0: Artificial intelligence or intelligence augmentation for safer process operation? AIChE Journal, e18475.