Aditya Johri, Andrew S. Katz, Junaid Qadir, Ashish Hingle
{"title":"Generative artificial intelligence and engineering education","authors":"Aditya Johri, Andrew S. Katz, Junaid Qadir, Ashish Hingle","doi":"10.1002/jee.20537","DOIUrl":null,"url":null,"abstract":"The recent popularity of generative AI (GAI) applications such as ChatGPT portend a new era of research, teaching, and learning across domains, including in engineering (Bubeck et al., 2023; Kasneci et al., 2023; Lo, 2023; Qadir, 2023). In this guest editorial, we discuss the potential impact of GAI for engineering education as researchers and teachers. We see this editorial as the start of a serious dialogue within the community around how GAI can and will change our practices, and what we can do to respond to these shifts. GAI is built on foundational models (FMs) that can be adapted to various other tasks, such as large language models (LLMs), and they operate by learning from many examples and becoming very good at predicting the subsequent probable output or output sequence. Given the abundance of digitized data, they can quickly learn a wide range of topics and respond to user queries almost instantly. Whether engineering a new software application, writing a code snippet to analyze data, designing a product, or composing a cover letter for a job application, GAI users can leverage the power of LLMs to generate outputs that meet their specific needs (UNESCO, 2023). The ability to learn a skill and adapt it to new contexts is a capability that humans have excelled at for a long time. Some would even argue that the competence to learn original things in new environments to tackle novel problems, and teach it to others, is one of the most unique characteristics of our species (Tomasello, 2009). To assist us in this process, we also have the capability to continually create tools and techniques, another distinct trait of humans and central to the engineering profession (Johri, 2022). What, though, is the potential and limit of developing tools and technologies that can mimic and even go beyond what we have conceived of as human intelligence? What potential consequences do technology that can generate novel outputs have for society, especially education in terms of both benefits and harms (Bommasani et al., 2021; Farrokhnia et al., 2023)? What implications does this have for engineering educators (Johri, 2020)? While we discuss how GAI shapes research and teaching practices within engineering education, we recognize that there are additional implications for the use of GAI for self-motivated and sustained learning initiated by learners on their own. That topic is beyond the scope of this editorial and discussed in some detail in the Menekse et al.0s guest editorial in this issue.","PeriodicalId":50206,"journal":{"name":"Journal of Engineering Education","volume":null,"pages":null},"PeriodicalIF":3.9000,"publicationDate":"2023-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Engineering Education","FirstCategoryId":"5","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/jee.20537","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"EDUCATION & EDUCATIONAL RESEARCH","Score":null,"Total":0}
引用次数: 1
Abstract
The recent popularity of generative AI (GAI) applications such as ChatGPT portend a new era of research, teaching, and learning across domains, including in engineering (Bubeck et al., 2023; Kasneci et al., 2023; Lo, 2023; Qadir, 2023). In this guest editorial, we discuss the potential impact of GAI for engineering education as researchers and teachers. We see this editorial as the start of a serious dialogue within the community around how GAI can and will change our practices, and what we can do to respond to these shifts. GAI is built on foundational models (FMs) that can be adapted to various other tasks, such as large language models (LLMs), and they operate by learning from many examples and becoming very good at predicting the subsequent probable output or output sequence. Given the abundance of digitized data, they can quickly learn a wide range of topics and respond to user queries almost instantly. Whether engineering a new software application, writing a code snippet to analyze data, designing a product, or composing a cover letter for a job application, GAI users can leverage the power of LLMs to generate outputs that meet their specific needs (UNESCO, 2023). The ability to learn a skill and adapt it to new contexts is a capability that humans have excelled at for a long time. Some would even argue that the competence to learn original things in new environments to tackle novel problems, and teach it to others, is one of the most unique characteristics of our species (Tomasello, 2009). To assist us in this process, we also have the capability to continually create tools and techniques, another distinct trait of humans and central to the engineering profession (Johri, 2022). What, though, is the potential and limit of developing tools and technologies that can mimic and even go beyond what we have conceived of as human intelligence? What potential consequences do technology that can generate novel outputs have for society, especially education in terms of both benefits and harms (Bommasani et al., 2021; Farrokhnia et al., 2023)? What implications does this have for engineering educators (Johri, 2020)? While we discuss how GAI shapes research and teaching practices within engineering education, we recognize that there are additional implications for the use of GAI for self-motivated and sustained learning initiated by learners on their own. That topic is beyond the scope of this editorial and discussed in some detail in the Menekse et al.0s guest editorial in this issue.