{"title":"Artificial intelligence (AI) and process safety: Some cautionary observations","authors":"Walter Frank","doi":"10.1002/prs.12641","DOIUrl":null,"url":null,"abstract":"Artificial intelligence (AI) has become a vogue topic in the press, and descriptions of its potential impact range from apocalyptic to salvational. Interest in the topic will no doubt stimulate the search for applications to support both the technical and management systems aspects of process safety management. Within our industries, maintaining institutional memory and technical capability is made increasingly challenging by more frequent job movement among younger staff and the loss to the retirement of more senior staff. One would hope that AI could help fill the gaps caused by these factors. However, the author's sampling of current AI capabilities suggests that AI is not yet ready to do so. This paper provides some examples of errors and insufficiencies identified when seeking AI assistance in addressing process safety issues. It also suggests some existing challenges to better “training” of AI to support the needs of the process safety community. It concludes that caution should be applied, especially by less experienced personnel, when seeking AI assistance in addressing process safety–related technical matters.","PeriodicalId":20680,"journal":{"name":"Process Safety Progress","volume":"12 1","pages":""},"PeriodicalIF":1.0000,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Process Safety Progress","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1002/prs.12641","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"ENGINEERING, CHEMICAL","Score":null,"Total":0}
引用次数: 0
Abstract
Artificial intelligence (AI) has become a vogue topic in the press, and descriptions of its potential impact range from apocalyptic to salvational. Interest in the topic will no doubt stimulate the search for applications to support both the technical and management systems aspects of process safety management. Within our industries, maintaining institutional memory and technical capability is made increasingly challenging by more frequent job movement among younger staff and the loss to the retirement of more senior staff. One would hope that AI could help fill the gaps caused by these factors. However, the author's sampling of current AI capabilities suggests that AI is not yet ready to do so. This paper provides some examples of errors and insufficiencies identified when seeking AI assistance in addressing process safety issues. It also suggests some existing challenges to better “training” of AI to support the needs of the process safety community. It concludes that caution should be applied, especially by less experienced personnel, when seeking AI assistance in addressing process safety–related technical matters.
期刊介绍:
Process Safety Progress covers process safety for engineering professionals. It addresses such topics as incident investigations/case histories, hazardous chemicals management, hazardous leaks prevention, risk assessment, process hazards evaluation, industrial hygiene, fire and explosion analysis, preventive maintenance, vapor cloud dispersion, and regulatory compliance, training, education, and other areas in process safety and loss prevention, including emerging concerns like plant and/or process security. Papers from the annual Loss Prevention Symposium and other AIChE safety conferences are automatically considered for publication, but unsolicited papers, particularly those addressing process safety issues in emerging technologies and industries are encouraged and evaluated equally.