. ©American Society for Engineering Education, 2025Experience Report: Reflections on Teaching Ethics Unethically [evidence-based practice,DEI]AbstractEngineering ethics education is essential for future graduates, yet it is often seen as a secondary‘complementary’ study, it is pushed to be more ‘engineering-like’ through focusing onquantitative methods, it is reduced to ‘rules and codes’, and assessment of engineering ethicsremains a mysterious process that people are willing to ignore problems within. In this paper, Iprovide an autoethnographical case study of my experience in my first year of teachingengineering ethics, where I engaged as an observer with another instructor’s content. I provide anoverview of the course, and present four main
’. ©American Society for Engineering Education, 2025Exploring engineering students’ understanding of their social responsibilitythrough a living library of ethics case studiesIntroductionEthics education is increasingly recognized as a crucial component of the undergraduateengineering curricula. Nonetheless, many engineering students show reluctance or outrightdisengagement when exposed to ethical issues [1] [2]. Traditionally, the engineeringcurriculum privileges technico-scientific knowledge, seeing it divorced from ethics andsocietal considerations, and relegating ethics tends to standalone courses or ancillary topicswithin broader coursework [3], [4]. This hierarchization of disciplines reflects a deeper‘depoliticization’ of engineering programs
diminished potential fornegotiating with its outputs [2]. The implementation of collaborative and reflective learning has thepotential to engage students with facets of ethical awareness that go along with algorithmic decisionmaking – such as bias, security, transparency and other ethical and moral dilemmas. However, thereare few studies that examine how students learn AI ethics in electrical and computer engineeringcourses. This paper explores the integration of STEMtelling, a pedagogical storytellingmethod/sensibility, into an undergraduate machine learning course. STEMtelling is a novel approachthat invites participants (STEMtellers) to center their own interests and experiences through writingand sharing engineering stories (STEMtells) that are
collaboratively written codes to analyze individual media case studies,identifying areas where their frameworks failed to address critical ethical dilemmas. This createdan opportunity to engage with ethical reasoning in a playful yet rigorous format, encouragingdeeper reflection and prompting revisions of their codes to better account for ambiguity,complexity, and edge-case scenarios in biomedical design ethics.Assignment DescriptionStudents were challenged to create a discipline-specific code of ethics tailored to biomedicalinnovation. Rather than simply summarizing or quoting standard documents, they were expectedto synthesize core concepts from foundational ethics texts with their own interpretations. Theassignment sequence began with a lecture series
Engineering & Materials Science Department Duke UniversityAbstractThe increasing use of artificial intelligence (AI) in recruitment, particularly through resumescreening algorithms, raises significant ethical concerns due to the potential for biaseddecision-making. This case study explores these issues by developing a synthetic datasetmimicking the Amazon hiring tool controversy, where biases in training data led todiscriminatory outcomes. Using artificial resumes that reflect a diverse applicant pool, studentstrained and interacted with a machine learning algorithm, which, despite excluding explicitdemographic information, exhibited biases against underrepresented groups. This exercisehighlights
. Students should be able to articulate the challenges and harms that GAI tools and LLMs can cause, and acknowledge how these percolate into their usage in the UCD process. 2. Students should be able to incorporate GAI tools and LLMs into the UCD process, and make informed decisions on whether using such tools at any given point is appropriate. 3. Students should gain practical experience working with GAI tools and LLMs within various stages of the UCD process, and be able to reflect upon the efficacy (or lack thereof) of such usage. 4. Students would develop an understanding of the growing body of research on GAI tools and LLMs, and gain insights on the direction of the field. Each class period
presented by Rotolo, Hicks, and Martin[18] as possessing “radical novelty, fast growth, coherence, prominent impact, and uncertainty andambiguity.” In another framing of emerging technologies such as CRISPR/Cas9 and AI,Veluwenkamp et al. [19] discuss “socially disruptive technologies” that require us to reflect on thenomenclature and significance of novel technologies that can lead to ethical design practices. BothRotolo et al. and Veluwenkamp et al. emphasize the importance of discussing emergingtechnologies at all stages of innovation to prepare for an ethical future of responsible innovation,development, and deployment. In this paper, I will refer to emergent technologies as novel, underexplored, and rapidlyevolving technologies that are
, treating them asperipheral to the core responsibilities of engineers [1], [2].The foundations of engineering ethics can be traced back to early professional codes developedto address the responsibilities of engineers in ensuring public safety and reliability. For example,the Canons of Ethics by the American Society of Civil Engineers emphasized technicalcompetence, safety, and accountability [5]. Over time, engineering ethics evolved to includebroader societal concerns, such as environmental stewardship during the environmentalmovements of the 1960s and 1970s. Frameworks like sustainable design and corporate socialresponsibility emerged, reflecting a growing recognition of the interconnectedness betweenengineering practices and societal impacts [4
topreserve critical thinking and foundational writing skills. Both groups called for clearerinstitutional policies and structured guidelines for the ethical use of AI tools in educationalcontexts.The findings underscore the need for a balanced and proactive framework to leveragegenerative AI’s benefits while safeguarding educational integrity. Key recommendationsinclude: (1) establishing clear institutional policies on permissible AI use; (2) developing AIliteracy modules to foster critical engagement; (3) implementing process-oriented assessmentmodels, such as version history reviews and reflective writing logs, to emphasize students'intellectual contributions; (4) promoting active faculty involvement in guiding ethical AI use;and (5) adopting
complete an anonymousonline survey, followed by two reminders. Forty-five students completed the survey. Inthe second phase, the same cohort was invited to participate in semi-structuredinterviews, with three reminder emails sent. Twelve students participated in one-hourZoom interviews. Participation in the survey and interviews was voluntary, and informedconsent was obtained in both cases.Data CollectionThe survey collected quantitative data on ethical perceptions and experiences, while theinterviews provided qualitative insight into students’ reflections on ethics in research andpractice. All interviews were conducted via Zoom, recorded with consent, andtranscribed for analysis.Data AnalysisQuantitative data were analyzed to detect trends in
2010 to 2024, we manuallyannotated the true set of violated articles for each judgment and compared them to theLLM-extracted sets. The results were as follows: • Jaccard Accuracy: 91.88% • Precision: 91.88% • Recall: 100.00% • F1-score: 92.89% These results demonstrate that the LLM-based extraction pipeline performs with highfidelity. The perfect recall indicates that no relevant article references were missed (i.e., nofalse negatives), while the high precision reflects relatively few false positives. False positives stemmed mainly from: (1) Ambiguity in articles under suspended proceedings (2) Misinterpretation of restated charges during sentencing Despite these edge cases, the evaluation supports the reliability of the LLM
. Wechose to not rely on an ethical framework for reference, because we have found thatmany students have interpreted ethical frameworks in absolute terms.The exercise began with a briefing about the differences between ethics and morals, withexamples of typical moral themes, followed by individual reflection about what thestudents knew about themselves. The participants were then assigned to ad-hoc teams inorder to compare their moral priorities to those of other team members. Finally, eachteam formed a set of moral priorities for their own hypothetical engineering company.In order to assess the outcomes of this activity, we sought to answer the followingquestion: How did this exercise bring out multiple competing moral standpoints and
examples of ethics instruction identified by Walling [2] include: • Berne and Schummer’s use of discussion prompts to engage students in a discussion of the ethical implications of nanotechnologies featured in selected science fiction publication [3]. • McQueeney’s use of real-world business dilemmas to prompt students to write their own personal responses, writing which then serves to prompt class discussion [4]. • Johnston’s of ethical dilemmas for students to analyze in both writing and an oral presentation. In addition, students were asked to log or journal their developing ethical thoughts and concerns as they reflected upon the dilemma [5]. • Brummel et al’ s use role-play scenarios to teach
favorof there being multiple cognitive schemas available to a person depending on the specificsituation they are considering, although there can be a preferred schema. Despite the shift intheoretical frameworks, the DIT remained a primary assessment tool for studying moralreasoning, although the interpretation of results changed.The original DIT required test takers to read six stories concerning moral dilemmas and then rateand rank items related to the stories. In the 1990’s, the DIT was revised, producing the DIT-2,with new stories that reflected the changing social context [2].The original DIT used a numerical index, the P-score, that measured the percentage of post-conventional responses to a moral dilemma. The DIT-2 also uses the P-score
letter in thisacronym stands for a different stage/component of the meeting. These includes (B) bridging in,presenting findings or introducing an activity to pique interest in the topics that will be coveredin the meeting, (O) introducing objectives, informing participants what they will get out of themeeting, (P) pre-assessment, learning what participants know about the topics that will becovered, participatory learning, guiding participants to actively reflect as topics are introduced,and post-assessment, learning what participants understood about the topics covered, and (S)summarize the meeting, reminding participants what the objectives were and how these werefulfilled.Curriculum contentsThe workshop includes contents related to seven
judgments. ABET CE program criteria also specifies that the CE curriculum mustinclude the application of the American Society of Civil Engineering (ASCE) code of ethics toethical dilemmas. VMI's approach aims to embed ethics within both the curriculum and thebroader educational experience.Beyond a traditional CE curriculum, discussions of ethics arise in CE courses, LeadershipEducation and Development (LEAD) programs, and the Reserve Officers' Training Corps(ROTC) training, allowing students to reflect on the ethical implications of their engineeringchoices. Furthermore, in an extracurricular capacity, VMI's Honor Court further promotes aculture of integrity and accountability among students. Evidence regarding the extent to whichVMI's curriculum
, narrative-based game that immerses students in real-timeethical dilemmas. By placing students in authentic, problem-solving situations, Mars!encourages ethical reasoning that reflects the complexity and ambiguity of professional decision-making, rather than requiring students to apply pre-existing ethical frameworks in a detached,theoretical manner.Playful Learning and Stealth AssessmentThe integration of game-based learning into ethics education builds on research suggesting thatplayful environments encourage deeper engagement and more authentic decision-making [14].Narrative-driven games like Mars! provide students with interactive, immersive experiences thatrequire them to make ethical choices within realistic, high-pressure scenarios, rather
.” (translated with deepl) [1: p.74].In the general discussion, this requirement is reflected, for example, in the concept of the t-shaped engineer, whose strength is seen in the great variety of interdisciplinary skills, which,in addition to mastering foreign languages, include cultural and communicative skills. In addi-tion, young engineers are expected to think systemically and holistically, as well as to be ableto critically reflect on their own actions [2], [3]. A critical examination of the concept of the t-shaped engineer and a literature review in the context of the ASEE can be found in [4].The aim of these approaches is to lay a foundation for a technology and product developmentprocess that takes into account the non-technical and non-economic
thehorizontal axis, I draw on functionalist notions of professional status that depend on a specializedknowledge base, commitment to public good, and self-regulation. The disciplinary training boxand ethical box are to the right because they reflect two of the three professional statusrequirements (specialized knowledge base and commitment to public good), while themanagerial box is on the left since employers, clients, and government policies restrict the extentto which an individual engineer or even engineering as a profession may self-regulate. In thenext two paragraphs I use sociological theories to blur the seemingly static nature of myexplanation.The overarching theoretical umbrella for my analysis is based on Larson’s notion
purpose of this principle is to promote transparency and provide sufficient information to users tomake informed choices. The fifth principle, Human Alternatives, Consideration, and Fallback, emphasizesthe right of individuals to opt out of automated systems in favor of human alternatives where appropriate,while highlighting the importance of accessible, and effective fallback mechanisms. It places theaccountability of decision to humans overseeing the working of said systems.The AI Bill of Rights is not a legally binding document and the values reflected in it are not enforced law.Instead, it is a guideline that can assist in the development of policies and practices that protect civil rightsand uphold democratic values in the age of AI systems
. As a survivor of violence healing among other survivors, I discovered emotionshave a factual component and reason for an individual, something to process in order to respondafter reflection, not react impulsively, and are important to consider in the design of publicservices. In my graduate study at Syracuse University’s Maxwell School for Citizenship andPublic Affairs I learned that public policy success depends upon “thick” participation (definedlater in this paper), where qualitative research is an example. “Indifference towards people and the reality in which they live is actually the one and only cardinal sin in design.” — Dieter Rams, German industrial designerpublic policy failures and relevance of qualitative
education and career phases.All 38 HEEE invitees were asked to complete the pre-event survey, and we received 24responses, reflecting a response rate of about 63%. Given the nature of the event and invitationlist, all participants had some level of interest or experience related to promoting or studyingengineering ethics in university or workplace settings. Of the 24 respondents, 14 were mainlyaffiliated with academic institutions, including faculty, instructors, staff, and graduate students.Another seven were employed in or recently retired from the private sector, and two heldpositions in the public sector. Disciplinary backgrounds and affiliations ranged widely amongthis group, including individuals with engineering and non-engineering degrees
, stories, by design, are full of interesting dialogs, different characters, plot twists,and cultural elements, making them quite interesting and engaging. This aspect was mentioned bystudents repeatedly in the post-survey and students also expressed they enjoyed the virtue andethics teaching modules in the anonymous end-of-course evaluations (administered by theuniversity). Many students also noted the stories helped contextualize the virtues and gave themgood examples that they can reflect on when they need to make ethical decisions in their futurecareer. Stories from traditional culture have the added benefit of teaching students interculturalcompetencies (i.e., the ability to function effectively across cultures, to think and actappropriately
andminorities in research despite knowing that they exhibit different symptoms and have risk factorsdifferent to mainstream groups. Some facial recognition software has shown a bias in favor oflighter skin tones. If AI is trained on electronic health records, it is building on only people whocan access healthcare and is perpetrating any limits that are included those records. Health-related AI data need to represent a wide range of social and economic backgrounds; otherwise,algorithms will merely reflect the pre-existing biases of society when faced with situations thatinvolve ethical and social complexity. For example, if poor patients’ health conditions are foundnot to improve after chemotherapy, machine learning algorithms might recommend against
incentives to engageand retain users. The prospect of profit serves as a catalyst for tech industry growth, promptingus to pause and reflect on not just “what” products are being developed, but also “why” and“how” they are being created, as well as considering “who” is using those products.Undergraduate coursework offers an ideal setting in which to incorporate ethical andpsychological principles into engineering training. Traditionally, academic programs haveoffered diverse classes and training methods with focus on requisite technical and proceduralskills for innovation. There are varying treatments of how ethical and psychological conceptsare integrated in engineering and computer science programs where technology products arebuilt as part of