Boulder. Scott’s research relates to accessible and inexpensive engineering equipment for laboratory education. ©American Society for Engineering Education, 2023 Artificial Intelligence Solutions for Digital DesignAbstractAccessible artificial intelligence platforms, especially ChatGPT, are now available to solveengineering questions. Here we evaluate this tool for finite state machine construction in Python.With well-guided queries, ChatGPT built sensible code that implements a microwave ovencontroller for hardware integration. However, to leverage ChatGPT user knowledge of theprogramming task was necessary, which included schematics, input, and output delineation, anddebug expertise.Special Note
, chaired nine conferences including 2009 ASEE/PSW and 2015 ASEE/PSW and three USPatents. ©American Society for Engineering Education, 2023 Student Use of Artificial Intelligence to Write Technical Engineering Papers – Cheating or a Tool to Augment LearningAbstractConsiderable concern has emerged over the potential use of AI tools by students for completingassignments in their classes. Reactions in academia have been mixed, with some describing suchuse of AI tools as “cheating” while others compare it to the use of calculators and see it as theimpetus for enabling deeper learning by students. To analyze some of these issues, the recentlyreleased AI tool ChatGPT was used to respond to actual Discussion
students learn better and fasterwhen paired with high-quality learning materials and instruction. But due to the array ofavailable specializations in industry categories, selecting the best fit for their interests is a bigchallenge for engineering students. This paper focuses on using AI to help students choose theirengineering program.Using ChatGPTChatGPT is a chatbot launched by OpenAI in November 2022. It is built on top of OpenAI'sGPT-3 family of large language models and is fine-tuned with both supervised andreinforcement learning techniques. We tried to use ChatGPT as a trained AI system to answerthis question, “How ChatGPT helps students to select their program.”ChatGPT answer was:“ChatGPT is a large language generation model developed by
., vertical versus lateral reading), but also toevaluate and incorporate sources written by non-subject-matter experts (e.g., how one mayutilize research journalism and news reporting versus scientific and technical publications).Natural language processing (NLP) models such as ChatGPT are also included in the sourceevaluation exercises, reflecting emerging concerns about how they will affect research andwriting.This paper argues that source evaluation is a skill that must be taught in all STEM classrooms;the stakes for society of producing STEM graduates with a poor research foundation are simplytoo high. Just as STEM students learn to flex their critical-thinking skills to make reason-basedprofessional judgments, they can apply those same critical
attention inthe business and public sphere with the release of models like ChatGPT [4] and DALL-E [5],robust applications within the field of engineering education remain are still emerging [6]. Aspart of the recent popularity of large language models (LLM) there have been increasingconcerns about the ethical ramifications in educational and industry settings. In their analysis ofthe practical ethical dangers of ChatGPT Zhuo et al. [7] outline areas of concern for LLMs as agroup; the risk inherent in small models propagating with increased scale, potential biases withinmodel training data, and the ballooning size of LLMs computational requirements. Theseconcerns limit the number of practitioners that are willing to adopt ML, NN, or LLM tools
be better. Everyone is used to that now and this ChatBox technology seems a bit outdated (e.g. people are using Google and this is Yahoo). Incorporating AI into it might make it a lot better.” “Ask follow-up questions after asking ChatGPT, and somewhat integrate the chatbot answers with the response acquired from ChatGPT (record it in database or something so that when the same or a similar question is asked again the bot is able to provide answers relevant to the course)” 9Overall, the preliminary finding based on the initial prototype of the chatbot was promising. Thechatbot made accessing information much easier for students and other users
[of the students] and holding it’ [3]. He goes on to contend that our task as professorsis not unlike that of a commercial for a soft drink or other product- what differs, is what professorsdo with the attention once they have it.Stories are not just a means to record historical events but can be a powerful way to teach lessons.Jesus Christ, arguably one of the most influential individuals in history relied heavily upon the useof stories (parables) and metaphor to teach His disciples, who in turn have passed these storiesdown for over 2000 years.Not into religion? If your faith lies instead on the future of Artificial Intelligence, then considerthat ChatGPT lists “storytelling” as its number 2 strategy when asked how to make engineeringvideos
includegenerative AI models such as the text-to-text model, chatGPT [9], GPT-4 [9], and others. LLMsuse a transformer model architecture instead of a CNN and the transformer architectures arecurrently being explored for use in computer vision applications. Models such as chatGPT (andothers) have also proven useful for programming code generation and productivity enhancement.These LLMs are of growing importance but are outside the scope of this project and paper.2. Deep Learning Curriculum and ProjectsThe specific goal of this project is to design and implement an instructional 7-week coursemodule to introduce deep learning and computer vision with a project-based orientation. Asmentioned, the target course is a senior-level engineering design course in
integrated into ScribeAR butother integration projects are possible. For example, ESPnet [21] is also a common speech to textplatform that is an end-to-end speech processing toolkit. It includes various applications such asspeech recognition, text-to-speech, speech translation, and speech enhancement.Recent advances in Large Language Models will also provide new opportunities for inclusiveconversational approaches. For example, a student project might use the new ChatGPT API thatas of May 2023 is now available as part of Microsoft Azure cloud services, to providesummarization or other textual transformation of a transcript [24]t.8. AcknowledgmentsWe thank the VR@Illinois program and the Department of Physics Graduate Office at theUniversity of
stand-alone EC course.This trajectory is set to change rapidly with the rise of interdisciplinary, remote engineeringteams, increasingly visually-focused publication modes [25], and ChatGPT [26] and other AI-powered writing tools. It is beyond the scope of this paper to detail communication-supportingapproaches to integrating AI- and non-AI-powered tools in the EC course context [27]; however,selected emerging apps show clear promise for students for visual and presentation contexts(Tome.ai, Orai) [28-29] and written contexts (WordTune, ChatGPT) [30].The New EC CurriculumIdeally introduced in the second year of engineering curricula, an EC course is able to scaffoldand develop authentic communicative capacity grounded in students’ interests and
become whistleblowers be taken seriously, or not [7]; • Is the Amazon machine-learning algorithm used for recruiting discriminatory against women, or not [8]; • Should controversial public people be banned from Twitter and other social media platforms, or should the First Amendment protect them [9]; and • Should ChatGPT be embraced in school settings, or should it be banned [10], [11]. Acknowledging the relevance of ethics in CS education is not a novelty. In 1972, ACMreleased and adopted the first Code of Professional Conduct [12], with its last revision releasedin 2018 [13]. Discussions of professional and social responsibility in CS education have beenpart of professional forums for decades [14]. The 2017-2018
John Searle’s Chinese Room Problem [6]. [7]This remains true with ourinteractions with large scale natural language AI such as the ChatGPT that has generated somuch current buzz. In this case it is not a brute force lookup but a prior machine learning fromprior examples of natural language. That does not mean that these language AI can ‘understand’.It could be that aspects of language such as humor or lying or subtext generally are the frontierwhere Chatbots fail [7].Whether programming is sequential, parallel, object-oriented or any other variant, the branchingof decision-making is constrained by what the programmer envisioned. There is no adaptation toevents or the environment except by the conditionals that the human programmer had
, compellingly titled“ChatGPT: Bullshit Spewer or the End of Traditional Assessments in Higher Education?”, the authors discuss the threatof ChatGPT to academic professionals and provide recommendations to them in the face of the growing expansion ofpowerful natural language models. They conclude with the following: “… we believe that major changes to traditionalhigher education assessments such as essays and online exams are in order to address the existence of increasinglypowerful AI, unless universities want to be akin to driving schools that teach [horse riding]” [18]. This paper does not long consider language models that can be used to write student assignments; it is mentionedas an area of tangential concern to educators. The primary focus of
, and novelty. First,artists are upset because the datasets include their unauthorized copyrighted material. The Large-scale Artificial Intelligence Open Network-5B (LAION-5B) makes one of the larger datasetrepositories which holds 5.85 billion usable image and text pairings [9]. Then product vendorsaccess the dataset and deploy diffusion models to train the AI. The text-to-image generators donot attribute the original work [10]. Second, computers do too much of the work and not theartist. Using minimal effort, Ammar Reshi engages ChatGPT to write and AI art to illustrate thechildren’s book Alice and Sparkle [11]. Third, AI art is novel, and scholars have not vetted andresearched the curriculum for its impact on architecture education and
videos in a more efficient way. The trend is similar to the findingsmentioned in [19], [20].Figure shows how many hours students spent on their weekly studies. The majority of studentsread and studied 3-5 hours each week out of class, which was consistent with the ideal workloadreflected by the majority from Figure . ChatGPT [21] was used to analyze students’ responses totwo open-ended questions, Q6 “What could you do to be more successful next week?” and Q7“What could your instructor do to help you be more successful next week?”.Figure 7 Shows the sentiment results of Q6, “What could you do to be more successful nextweek?”. After analyzing students’ comments, the most frequent topic is "time management",indicating that students feel they need
receive instant feedback about their score but do not have access to thequestion or their answer after they submit it. Students can re-attempt these exam topics infinitetimes; getting a (hopefully) different question each time.Students are free to access their notes and course materials while answering these questions andare encouraged to write and run test code as well. They are allowed to access the Internet ingeneral, however, they are forbidden from using Internet chat sites such as stackoverflow.com orquestion answering sites such as Chegg, CourseHero, or ChatGPT or from communicating withother people using any medium.Exam questions give the students 30 minutes to do problems that would typically take them 5-10minutes on a standard exam
classes. Implementing these methods to civilengineering problems in major requirements and electives can be mentioned to allow students torecall these concepts throughout their undergraduate career. High technology will only becomemore available, such as ChatGPT, and instructors and lecturers must prepare students for aworkforce that can use these methods for their own professional benefit. Section 2 provides moredetails on how universities can consider ABET accreditation when making high technologychanges in their curriculums.2. Accreditation and continuous improvementCivil engineering programs seek accreditation in order to ensure their program meets definedstandards of quality set by an accreditation organization. Accreditation also ensures
not a definitive and exhaustive list, it is a good starting point andcan be modified to suit the specific needs and context of the partnership. (Note that the rubricwas constructed with the assistance of ChatGPT. Putting a collection of related ideas into astandard format, like a rubric, is one of the tasks that AI seems to do well. We share thisinformation when we share the rubric as an incentive for potential collaborators to improve it.)Workshop PlanThe IEC has developed a broad and very challenging vision of enabling MSIs and their students,staff and faculty to more fully become part of and contribute to the ECE enterprise. To realizethis vision, this workshop brought together IEC Core MSI members with the heads of other ECEdepartments
lives. Thisis especially true now, since the world is in the midst of a number of controversies dealing withbiased data sets for training of neural networks, ChatGPT unfair uses, or the Elon Musk’s call fora moratorium on AI development. Results from this research will be used as preliminary findings while planning large-scale regionalresearch activities related to AI that could be supported by NSF, Amazon Machine LearningUniversity or the Department of Education. A collaborative network consisting of localschoolteachers interested in AI and AI-active university professors will be created to furtherpromote and implement AI in the K-12 curriculum. Partnership modalities with the AI4K12organization will be investigated to improve AI literacy
to the question “Was there anything that you would take away fromthis experience and apply to future projects or group work? If so what?” in the survey wasanalyzed and validated in two ways. Salient themes were identified by the authors and thefrequency of those themes were tabulated to count the number of occurrences specific featureswere identified from students. Responses were coded into the following 8 themes:Communication, assigned roles, motivation, lab skills/course content, collaboration/teamwork,leadership, enjoyment, and delegation/group organization. In addition, the responses were inputinto an artificial intelligence natural language processing tool (ChatGPT, OpenAI) to identifythemes from responses in an unbiased manner. This
learning community of the course? Ex. Engage students with personal knowledge that can enhance class activities like students from various origins or countries who can discuss how climate change is affecting their homes YES or NO Are students required to demonstrate self-reflective processes in evaluating engineering in society? Ex. Require students to express and defend opinions on engineering issues in the news on a regular basis, like the effects of ChatGPT on education YES or NO Are there opportunities for students to demonstrate their ability to integrate multiple values into evaluation and decision making in an engineering context? Ex
students might help to gainmeaningful insight from students’ viewpoints to improve the AI certificate program. Finally, thelow number of student samples and institutional and regional effects may also be considered fora complete study in the future.6. AcknowledgmentThis work was funded and supported by the National Science Foundation (#2051113) andUSDOT Transportation Consortium of the South-Central States (TRAN-SET) (# 21-034, and#21-049) to YFJ.References[1] B. D. Lund and T. Wang, "Chatting about ChatGPT: how may AI and GPT impact academia and libraries?," Library Hi Tech News, 2023.[2] R. R. Murphy, Introduction to AI robotics. MIT press, 2019.[3] K. Siau and W. Wang, "Building trust in artificial intelligence, machine learning, and
Improving Decisions in Engineering Education Agents and Systems (IDEEAS) Lab, a group that uses multi-modal data to characterize, understand, a ©American Society for Engineering Education, 2023 Pushing Ethics Assessment Forward in Engineering: NLP-Assisted Qualitative Coding of Student ResponsesAbstractRecent headlines have featured large language models (LLMs), like ChatGPT, for their potentialimpacts throughout society. These headlines often focus on educational impacts and policies. Weposit that LLMs have the potential to improve instructional approaches in engineering education.Thus, we argue that as an engineering education community, we should aim to leverage LLMs tohelp resolve
degrees that werebenchmarked in more detail, 19 ‘engineering’ and ‘general engineering’ degrees required a lowerpercentage of technical coursework and offered a lower percentage of curricular choicecompared to 7 degrees that included the word interdisciplinary, integrated, or multidisciplinary intheir name. A few programs require students to take the NCEES Fundamentals of Engineering(FE) exam prior to graduation. The AI-based program ChatGPT definitions of general,interdisciplinary, and integrated all emphasized breadth, multiple disciplines, and design, whilealso including the distinguishing factors of practical (for general) versus complex and innovative/novel (interdisciplinary and integrated), and the importance of social impacts (integrated
, iNaturalist, and Merlin Bird ID show the rise andregularity of crowdsourced data. The rise of public-accessible machine-learning-based productslike ChatGPT, Dall-e, and iPhone-unlocking facial recognition illustrate the emergence ofartificial intelligence (AI) across vast public sectors. AI is predicted to affect global productivity[3], to promote/expose problems in diversity, equity, and inclusion [4], to impact conservationand biodiversity monitoring [5], and to increase the ability to do climate monitoring andforecasting [2]. The 2020 Nature Communication study [2] suggests that AI will influence theability to meet all 17 Sustainability Development Goals (SDGs) set out in their 2030 Agenda [6].Vinuesa, et. al. illustrate and discuss how AI can
more learner-centered and focusedon formative feedback rather than summative evaluation [1].As the world of technology continues to advance, a shift and embracement of new assessmentmethods is appropriate and necessary. For example, as of early 2022, the New York Timesreported that universities are now having to change the way they are teaching and assessingstudents because of the widespread availability of A.I. Chatbots such as ChatGPT [12]. WithinIE at Minnesota State University, Mankato, differential methods of assessment have beenemployed for over a decade and we want to share our experiences with oral exams to supportothers in embracing the changing world, better preparing engineering students for their futurepositions.Overview of
Theory uses a formalized network diagraming convention to model environmentalsettings [13]. The network diagram consists of a Subject, Mediating Artifacts, Object, andOutcome. The Subject uses external (LMS, computer devices) and internal (plans, strategies)tools to complete an Object (milestone) thereby achieving a desired Outcome. The tools, alsoknown as Mediating Artifacts, are imbued with cultural, historical, and social significance.Mediating Artifacts influence the behavior of the Subject using them, and in turn, the largersocial environment the Subject inhabits. A simple example is shown in Fig. 1. A Subject(Student) is tasked with writing a report on “Activity Theory” (Outcome). The Student (Subject)uses ChatGPT, Wikipedia, and Google