Beyond the Hype: A Comprehensive Review of Current Trends in Generative AI Research, Teaching Practices, and Tools

1 Introduction

• The report distinguishes teaching students to use GenAI to write code from using GenAI to support course teaching, such as through help-seeking bots or code feedback.The authors present this distinction as useful terminology for describing classroom interventions.Source: §1 · p. 2

1.1 Goals

• The report combines a systematic literature review, an international educator survey, and interviews with educators, researchers, and GenAI tool creators.It also surveys software developers in industry and triangulates data sources for its results.Source: §1.1 · p. 2, §1.3 · p. 3

1.2 Contributions

• The report reviews research through May 23, 2024, surveys instructors about GenAI integration and student competencies, and examines educators’ use of GenAI and perspectives on learning outcomes.Its contributions also include the industry perspective on GenAI use, policies, motivations, and expected future competencies.Source: §1.2 · p. 3

1.3 Structure of the Report

• The report presents literature-review methods and results, followed by educator and developer survey and interview methods and triangulated findings, then discusses validity, conclusions, and future work.The discussion of validity appears in Section 7, the conclusion in Section 8, and future work in Section 9.Source: §1.3 · p. 3

2 Systematic Literature Review: Methods

• The review asks how GenAI evidence in computing education can be summarized, how it is incorporated into teaching, and what motivates its use.Its focus is empirical evidence about student experiences with classroom interventions, extending prior work on student perceptions.Source: §2 · p. 3

2.1 Search String Construction

• The search combined terms for computing education, GenAI, pedagogy or teaching tools, and empirical methods across five databases.The databases were ASEE Peer, arXiv, Scopus, ACM Digital Library, and IEEE Xplore; the search cutoff was May 23, 2024, and 1,536 papers remained after duplicate removal.Source: §2.1 · pp. 3, 5

2.2 Paper Filtering Process

• The review retained empirical studies featuring GenAI in tertiary computing education and classroom interventions, while excluding work without human evidence, K-12-only studies, and short or non-research formats.Two reviewers screened titles and abstracts, advancing a paper when either reviewer judged it relevant; full-text reading could still exclude papers.Source: §2.2 · p. 5

2.3 Extraction

• Four researchers extracted data from the final 71 papers using a structured form.The review process included full-paper reading and a further eligibility check before extraction.Source: §2 · p. 3, §2.3 · pp. 5–6

3.1 Descriptive Statistics

• The 71 reviewed studies covered varied course and author contexts: 20 studied CS1, 26 upper-division courses, and the largest author groups were in the United States (34%) and New Zealand (14%).Most studies used mixed methods; the median participant counts were 52 for mixed-methods, 56 for quantitative, and 21 for qualitative studies. Most evaluations were unsupervised, and Python was the most common language. Most empirical papers were sufficient on four quality dimensions, though limitations were vague in 20% or excluded in 11%.Source: §3.1 · pp. 6–7
• The review identified many custom tools and classified 57 papers as using GenAI for a pedagogical purpose and 14 as teaching students about GenAI or a specific tool.Examples of custom tools included CodeAid, CodeHelp, WorkedGen, Tutor Kai, and CS50 Duck.Source: §3.1 · p. 7

3.2 How Is Generative AI Being Incorporated into Teaching?

• Positive findings were reported in 54% of studies using perceptual evidence, 69% using behavioral evidence, and 70% using both.One study with positive findings had an unclear evidence type, so the table’s totals sum to 70.Source: §3.2 · p. 8, Table 6 · p. 8
• General-purpose tools were most common (34/71); 32/71 studies used instructor-provided guardrails, and 49/70 did not report student guidance.Common uses included code writing, comprehension, hints, and learning resources.Source: Table 5 · p. 8, §3.2 · p. 8

3.3 What Are the Motivations Behind Incorporating GenAI Tools into Teaching?

• GenAI was incorporated for students in 59 papers, for teachers in 5, and for both groups in 7.The review tags the intended beneficiaries of the 71 studies as students, instructors, or both.Source: §3.3 · p. 8

3.4 Recommendations for Incorporating GenAI

• 55% (12/22) of studies found positive results with unguided general-purpose tools, versus 73% (19/26) with unguided instructor-guardrailed tools.With guidance, positive findings were 58% (7/12) for general-purpose tools and 67% (4/6) for instructor-guardrailed tools.Source: §3.4 · pp. 8–9, Table 7 · p. 9
• 86% (6/7) of studies on learning resources and 80% (8/10) on code comprehension reported positive results, compared with 58% (15/26) for code writing and 4/8 for hints.The authors characterize hint generation as an area where GenAI could still improve.Source: Table 8 · p. 9, §3.4 · p. 9

4.1 Prior Studies of Educators’ and Developers’ Perceptions

• Prior studies largely examined educators’ prospective adoption rather than actual integration, while fewer studies focused on developers; this study surveys both groups [154].Earlier interview research reported concerns about over-reliance and misuse, uncertainty about adapting courses, and appreciation of GenAI as student support alongside risks to learning.Source: §4.1 · p. 10

4.2 Research Questions

• The ten research questions cover course policies, teaching students to program with GenAI, instructor support uses, motivations, emerging tools, and perceived student outcomes.The study distinguishes teaching students to use GenAI for programming (RQ2) from instructors’ use of GenAI in course activities and workflow (RQ3).Source: §4.2 · p. 10
• The remaining questions address developers’ industry use, desired competencies, student equity, and GenAI’s future in computing education.Source: §4.2 · p. 10

4.3 Mixed-Methods

• Two surveys and interviews combine breadth and depth, and findings are integrated by research question rather than separated by method.The interviews include tool creators, educators studying GenAI, and educators using it.Source: §4.3 · p. 11

4.4.1 Educator Survey.

• Educator survey questions address concrete teaching approaches, assessments, changing programming competencies, and students’ expected GenAI tools and access costs.The educator survey branches according to whether instructors allow student use; the developer survey covers professional use, workflow impact, competency changes, and ethical concerns or harms.Source: §4.4.1 · p. 11, §4.4.2 · p. 11

4.6 Survey Data Analysis

• The analysis uses descriptive statistics for closed-ended survey items and thematic coding for open-ended responses; it includes 76 complete educator responses and 39 developer response sequences.Of the developer sequences, 29 were fully complete; two researchers independently coded a subset of open-ended responses and resolved disagreements through discussion.Source: §4.6 · p. 13

4.7 Interviews Method

• The interview study included 17 participants: four tool creators, five educators studying GenAI, seven educators using it, and one thoughtful non-user.Participants were selected based on the literature review; interviews were one-on-one over Zoom and limited to one hour.Source: §4.7 · p. 13, §4.8 · p. 14, §4.8.1 · p. 14

4.7.1 Interview Question Development.

• Interview questions were collaboratively refined for clarity and distinct coverage, then asked sequentially without being shared beforehand; recordings were transcribed and deleted after transcription under IRB requirements.Follow-up prompts sought further context, and interviews were kept to one hour; the transcript analysis involved multiple team members to support consistency and reliability.Source: §4.7.1 · p. 13, §4.8.3 · p. 14, §4.8.2 · p. 14

4.9 Interview Data Analysis

• Three team members jointly tagged one transcript to align on predefined categories tied to the research questions.Categories included Instruction-use, Instruction-teach, Tools, Policies, Motivations, Competencies, Outcomes, Future, Industry, and Equity.Source: §4.9.1 · p. 14

4.9.1 Initial Review and Consensus Building.

• After consensus, a researcher who was not the interviewer independently tagged the remaining transcripts and compiled categorized segments in Excel.The paper says this approach helped maintain objectivity and consistency.Source: §4.9.2 · p. 14

4.9.2 Individual Tagging and Data Compilation.

• Researchers summarized each segment, then grouped summaries by category, identified patterns and themes, and synthesized community thoughts.The summaries were intended to capture interviewees’ responses without losing context or main points.Source: §4.9.3 · p. 14, §4.9.2 · p. 14

4.9.3 Thematic Analysis.

• The analysis aimed to derive conclusions from interview data and understand community views across the topics covered.Source: §4.9.3 · p. 14

5.1.1 Educator Survey Demographics.

• Educator respondents represented North America, Europe, and Australia; many taught CS1, and their teaching experience averaged 15.8 years with a 14.5-year median.The survey included 76 educators; 31.6 % identified as female, 57.9 % as male, and 52 of 76 described their institution as a university granting graduate degrees. Respondents reported teaching 193 classes across varied computing courses.Source: §5.1.1 · pp. 14–15

5.1.2 Developer Survey Demographics.

• Developer survey responses came from mostly US- and Germany-based participants, most of whom identified as software developers working at large companies.The respondent counts varied by question: 18 reported country, 23 job title, and 27 company characteristics; 59.3 % of the latter group worked at companies with more than 500 employees.Source: §5.1.2 · p. 15

5.1.3 Interview Sample Demographics.

• Interviewees were selected for diverse experiences and perspectives, with deliberate outreach to K-12 teachers, non-CS computing professors, and thoughtful GenAI non-users.The candidate list drew on personal knowledge, the literature review, conference networking, and a mailing list; substitutions were made when selected candidates declined.Source: §5.1.3 · pp. 15–16

5.2.1 Educator Policies preventing, tolerating, or integrating GenAI.

• 77.6 % of surveyed educators did not explicitly disallow student GenAI use, 35.5 % integrated it, and 75 % believed software-creation skills had changed.Among educators seeking to prevent use, 36.0 % reported code reviews, 28.0 % assignment or exam changes, and 20.0 % appeals to social responsibility. [The passage reports 9 out of 25 for code reviews.]Source: §5.2.1 · p. 16
• Educators who integrated GenAI described refocusing programming skills, setting use and attribution expectations, demonstrating tools in class, and teaching effective use.Some respondents emphasized reading code and prompt design, while others reported no changes yet; five themes summarized 60 open-ended responses.Source: §5.2.1 · p. 16
• Assessment changes included more process-focused evaluation, proctored or oral exams, greater exam weighting, unchanged practices, and requests to disclose GenAI use.Thirteen participants reported more emphasis on proctoring, ten increased the weight of proctored exams, and nine reported no changes, with six planning changes.Source: §5.2.1 · p. 17
• Interviewees raised privacy concerns and described institutional bans, instructor restrictions, and policies allowing or encouraging GenAI under conditions such as explainability or locally hosted tools.Eight developers stated they did not use GenAI; one cited a current company ban, while other replies did not address company policy.Source: §5.2.1 · p. 17, §5.2.2 · p. 17

5.3 RQ2 and RQ3 – Instruction-use and Instruction-support

• Among instructors integrating GenAI, 74.1% expected both instructors and students to use it, and common purposes included teaching GenAI use and generating educational content.All 27 reported using standard industry tools; interviewees described feedback and help-seeking uses, bias activities, code critique and testing, and industry preparation.Source: §5.3 · p. 18
• Interviewees also described using GenAI to create lesson plans and emphasized discussing model bias with students.One instructor said students were surprised by examples of bias and stressed the importance of addressing it.Source: §5.3 · pp. 18–19
• Interviewees described integrating GenAI into upper-division and software-development courses as a professional tool, while using student-facing tools to support learning without interfering with it.One instructor framed GenAI as part of a software-development workflow alongside IDEs, debuggers, and linters.Source: §5.3 · pp. 18–19

5.4 RQ4-Motivations

• Educators who barred GenAI cited foundational programming and curricular concerns, while adopters pointed to career readiness, ethical responsibility, learning support, and keeping pace with technology.The motivations for barring tools also included student misuse or lack of skills, doubts about capability, and legal or organizational concerns.Source: §5.4 · p. 19
• Among reported reasons for not integrating GenAI, lack of educator resources was cited in 24 statements, ahead of students’ perceived skills and motivation, mentioned by 10 educators.Educators cited lack of time, skills, or didactic competency; student concerns included assumed prior GenAI use and insufficient skills to use it properly.Source: §5.4 · pp. 19–20
• Other non-integration reasons included doubts about usefulness (8 responses), external factors (6 responses), and other teaching-practice issues (7 statements).External concerns included institutional support, legal questions, and privacy; other issues included not teaching a suitable course or discussing GenAI without further integration.Source: §5.4 · p. 20
• Interviewees favored education-focused tools because general tools can give complete solutions, whereas students need hints and guidance that reduce over-reliance.They described a Socratic approach that guides students toward solutions rather than simply providing answers.Source: §5.4 · p. 20
• Participants saw GenAI as a way to provide timely help and tailored feedback at scale, including to learners without access to human teaching support.They also described using it to ease instructor workloads and broaden student success in computing.Source: §5.4 · p. 20
• Researchers studying GenAI in computing education were motivated to understand how it affects students, learning, and what introductory programming should teach.They also sought to identify ways to use GenAI’s affordances to help students learn.Source: §5.4 · p. 20

5.5 RQ5 – Tools

• All 27 surveyed educators used standard GenAI tools, while course-focused tools added guarded coding help, course-material grounding, student-learning analytics, and simplified feedback.Only 2 of 27 educators used tools they had customized themselves; educators also reported reduced office-hour use after deploying assistants.Source: §5.5 · pp. 20–21

5.6 RQ6 – Perceived Outcomes

• Educators described both risks and benefits: GenAI can enable cheating, undermine learning, and mislead students, but can also support higher-quality work and tutoring.Some instructors also anticipated fewer fundamental coding skills, more advanced work, and a growing role for conversational programmers.Source: §5.6 · p. 21

5.7 RQ7 – Industry Usage

• 56.5% of developers reported daily GenAI use, compared with educators’ 36.4% expectation; developers most often reported autocompletion, chatbots, code generation, and code completion.Among surveyed developers, 31 of 39 used GenAI professionally; code generation and autocompletion were each selected by 17 of 24 task respondents (70.8%).Source: §5.7 · pp. 21–22, Figure 7 · p. 23

5.8 RQ8 – Competencies

• 57 of 76 educators believed software-creation skills had changed, including 25 of 27 whose courses integrated GenAI and 32 of 49 whose courses did not.Source: §5.8.1 · p. 22

5.8.1 Educators’ Perspectives.

• Educators who saw skills changing emphasized code reading, higher-level problem-solving, code evaluation and debugging, and prompting; problem understanding ranked first.Some educators instead viewed core skills as unchanged but shifting in importance, particularly toward code evaluation; others said it was too early to know.Source: §5.8.1 · pp. 23–24

5.8.2 Developers’ Perspectives.

• Developers identified understanding concepts, using GenAI productively, recognizing its limits, evaluating its suitability and output, prompting, and carefully checking generated code as relevant competencies.Views on declining competencies varied: some cited memorized syntax, external-resource use, or frustration tolerance, while others saw no skills as less relevant.Source: §5.8.2 · p. 24
• Most developer comments favored critical GenAI use and learning programming basics without it, rather than relying on tools early.Recommendations ranged from resisting or rarely using GenAI to using it critically or to code faster; the last approach was an exception.Source: §5.8.2 · p. 25
• Educators emphasized reading, evaluating, and modifying generated code, alongside prompt engineering and task decomposition, while warning that students may lose foundational computing knowledge.They also highlighted explaining code and computational thinking as important competencies as coding from scratch becomes less central.Source: §5.8.2 · p. 25

5.9 RQ9 – Equity

• Survey responses on whether institutions served minority populations were too limited and uncertain for further analysis of differences in GenAI teaching practices.23.7% (18 of 76) said yes, while 30.3% (23 of 76) were unsure; one participant did not answer.Source: §5.9 · p. 25
• Interviewees warned that unequal financial access, differences in self-regulation, language and literacy barriers, and disability-related interaction challenges could let GenAI amplify existing inequities.Some educators noted possible mitigations, including institutional licenses and tools designed for broader access.Source: §5.9 · pp. 25–26

5.10 RQ10 – Future

• Educators foresee GenAI supporting grading, assessment creation, code analysis, and tailored feedback, while assessments may increasingly evaluate AI interaction and prompt engineering alongside essential skills.They stressed careful planning and continuous adaptation to preserve critical thinking and deep learning, and described this future as still unfolding.Source: §5.10 · p. 26

Changing Competencies

• 77 % of faculty believe software-writing skills have changed, while 36 % have changed courses to include GenAI; about 30 % actively integrate it and 75 % acknowledge it in class.The discussion describes this as a clear trend, not a total change.Source: §6 · p. 26
• Recognizing that LLMs can solve take-home assignments, faculty are allocating more grades to proctored assessments and creating more or new proctored assignments.Source: §6 · p. 26
• Educators describe a shift from code writing toward code reading, higher-level problem solving, architecture, evaluation, testing, and debugging, while course changes often emphasize prompting.The literature review found more positive findings for code comprehension than for hint generation or code writing.Source: §6 · p. 26

Reasons for Disallowing GenAI

• Educators who disallow GenAI cite their own perceived lack of skills and students’ need to learn programming basics first; whether that stance is pedagogically sound remains an open research question.The authors also identify as unresolved whether GenAI can help beginners learn programming and in what ways.Source: §6 · pp. 26–27

Gap between Educators’ Expectations and Developers’ Actual Use of GenAI

• Educators expected GenAI use for more developer tasks than developers reported, while developers reported using it for getting started, debugging, code cleanup, and avoiding documentation.Modifying code was the only task where developers’ reported use exceeded educators’ expectations.Source: §6 · p. 27

Developers’ Use of and Trust in GenAI Tools

• 79.5 % of surveyed developers reported using GenAI, and 81 % reported increased efficiency; developers also expressed divided trust and concerns about complex-task performance.The comparison survey cited in the discussion reported 63 % GenAI use and 83 % increased productivity; its trust findings included 41 % with some trust and 31 % with some distrust.Source: §6 · p. 27

Equity Concerns

• GenAI access may create an equity burden: a standard three- or four-year program at $20 monthly could cost about $1,000, while access and students’ tool use may vary.The authors note that educators’ expectations about free-version use do not establish whether students use paid versions or how that affects outcomes.Source: §6 · p. 27

Need for Educators’ Professional Development

• Some educators report limited familiarity with GenAI and fixed views about its benefits, and the authors argue professional development could address barriers to pedagogically meaningful and critical engagement.The discussion connects limited exploration with educators potentially being less aware of risks and unable to discuss them critically with students.Source: §6 · p. 27

7.1 Systematic Literature Review

• The review did not search every computing-education database; its cutoff may omit nearby publications, and arXiv searches used titles and abstracts only.The authors also note likely publication bias, which could inflate the number of included studies reporting positive results.Source: §7.1 · p. 28

7.2 Educator and Developer Surveys

• Survey respondents came from mailing lists and may be more engaged with computing education and GenAI than their peers; consent requirements may also have discouraged participation.Both surveys rely on self-reports, which may include exaggeration, omissions, or socially desirable responses; open- and closed-ended questions and interview triangulation were used to mitigate this.Source: §7.2 · p. 28
• Developer-survey respondents may not represent typical software developers, and the researchers lacked access to developers at large technology companies worldwide.The authors correlated results with other industry-initiated research to mitigate this concern.Source: §7.2 · p. 28

7.3 Interview Study

• The educator-interview sample may not represent the computing-education community because participants were recruited and interviewed in English.Interview responses also carry self-reporting risks, including subjective representation, exaggeration, omissions, or socially desirable statements.Source: §7.3 · p. 28

8 Conclusions

• The report addresses how instructors incorporate GenAI and how software-development competencies may change through a systematic review, surveys of educators and developers, and educator interviews.The educator survey covers teaching and using GenAI, tools, policies, motivations, student competencies, equitable learning conditions, and future perspectives; the developer survey captures professional use patterns and potential harms.Source: §8 · p. 28, §8.2 · p. 29

8.1 Systematic Literature Review

• Across 71 reviewed papers, GenAI was mostly studied in unsupervised settings, commonly for code writing, comprehension, hints, and learning-resource generation, often without student-use instruction.Custom tools often included pedagogical guardrails intended to make GenAI use more productive for learning.Source: §8.1 · p. 28
• Studies more often reported positive findings with custom tools that included instructor scaffolding when students lacked explicit guidance on using GenAI.The review recommends guiding students who use general-purpose tools or using custom tools with pedagogical guardrails.Source: §8.1 · p. 28

8.2 Educator and Developer Views

• Educators interviewed described both cheating and exam unpreparedness, alongside responsible student use that can support greater accomplishment, especially in introductory courses.They also valued help-seeking tools that let students ask questions without the social risk of asking in front of peers.Source: §8.2 · p. 29
• Interviewees described computing competencies as being in flux, with some previously required competencies potentially giving way to new ones as education adapts to GenAI.The report presents this as instructors’ perspective rather than an established change in competencies.Source: §8.2 · p. 29

9 Future Work

• The authors propose repeating the review, surveys, and interviews in other domains to compare how educators integrate and perceive GenAI..Source: §9 · p. 29

Key takeaways

• 80% of surveyed developers use GenAI professionally, while 30% of educators report integrating it into classes and 22% explicitly disallow its use.The report presents these figures as evidence of differing levels of adoption among developers and educators.Source: §9 · p. 29
• Educators are increasing exam weighting, exploring invigilated formats such as oral exams, and emphasizing learning processes over answer correctness..Source: §9 · p. 29
• 75% of surveyed educators believe programming competencies have changed; they prioritize problem understanding, code reading, and decomposition, while developers also identify critical evaluation and prompt writing.Educators also point to testing, debugging, and understanding code as increasingly important skills.Source: §9 · p. 29
• The report recommends guiding students using general-purpose GenAI or using custom tools with pedagogical guardrails, while training educators and students to recognize tool limitations and challenges.Educators who did not integrate GenAI cited lack of time, skills, pedagogical training, support, or resources.Source: §9 · p. 29

9 Future Work

• Future tools could support learning, scaffolding, mastery learning, and debugging, while human tutors and educator training remain relevant to GenAI-supported teaching.The authors raise questions about effective teaching and instructor preparation as GenAI tools adopt teaching methods such as Socratic questioning.Source: §9 · pp. 29–30
• The report calls for regularly examining GenAI’s effects on the job market and graduate qualifications and aligning education with recent developments.It specifically raises the possibility of fewer entry-level software-engineering jobs.Source: §9 · p. 30

• The opening questions ask whether instructors disallow or incorporate GenAI, whether they believe software-creation skills have changed, and which skills matter most.Respondents are also asked how often professionals use GenAI and which tasks they use it for.Source: A Educator Survey Questions · p. 36
• The survey asks about professional GenAI-use frequency and tasks, reasons for disallowing student use, and steps instructors take to prevent it..Source: A Educator Survey Questions · p. 36
• For a recent GenAI-influenced course, instructors report course size, who uses GenAI, and how students access it, including public, paid, or custom tools..Source: A Educator Survey Questions · p. 36
• Course questions cover tool types, how GenAI is incorporated, reasons for integration, and whether learning objectives changed.Possible uses include feedback, grading support, teaching GenAI, and generating educational content.Source: A Educator Survey Questions · p. 36
• Instructors who have not integrated GenAI are asked why, while all respondents describe changes to teaching and assessment approaches..Source: A Educator Survey Questions · pp. 36–37
• Demographic questions cover country, institution, course area, teaching experience, gender, and willingness to be interviewed..Source: A Educator Survey Questions · p. 37

Use and practice

• The survey screens for professional GenAI use, then asks users about frequency, tool types, work examples, and the tasks for which they use the tools.Listed tasks include code generation, modification, autocompletion, debugging, documentation, and test-case generation.Source: B Developer Survey Questions · p. 37

Perceived effects

• Respondents rate GenAI’s usefulness, its effect on development efficiency, and its harmfulness, with an open question about harmful situations and their effects.Source: B Developer Survey Questions · p. 37

Competencies and advice

• The questionnaire asks whether required professional competencies have changed, what new or less relevant competencies respondents see, and what advice they would give novice programmers.Source: B Developer Survey Questions · pp. 37–38

Nonuse

• Nonusers are asked why they do not use GenAI, including company restrictions, doubts about coding help, and ethics concerns such as privacy.An open response invites elaboration, and nonusers are also asked for advice to novice programmers.Source: B Developer Survey Questions · p. 38

Demographics

• Demographic questions collect participants’ country, job title, and company type, with options ranging from startups to government and research institutes.Source: B Developer Survey Questions · p. 38

Tool and deployment

• Creators describe their GenAI tool and its users, reflect on deployment and modifications, and discuss institutional privacy and security policies.Source: C Tool Creators Interview Questions · p. 38

Outcomes and learning objectives

• The interview asks what outcomes tool use has produced, what data are collected, and which learning objectives the tool aims to reinforce.Source: C Tool Creators Interview Questions · p. 38

Expertise and future direction

• Creators discuss educator expertise needed to deploy and use the tools, future development plans, and views on AI’s role and needed tools in CS education.A final prompt invites any other discussion of GenAI.Source: C Tool Creators Interview Questions · p. 38

Research landscape

• Interviewees describe the categories and landscape of GenAI research in CS education and what they hope studying the area will reveal.Source: D Educators Studying GenAI Interview… · p. 38

Use, risks, and skills

• Questions cover technologies educators use, ways to prevent negatively perceived aspects, inequities in access or use, and changes in skills or competencies since LLMs emerged.Source: D Educators Studying GenAI Interview… · p. 38

Future direction

• Researchers are asked where they think GenAI in CS education is heading next.Source: D Educators Studying GenAI Interview… · p. 38

Course use and communication

• Educators identify the courses using GenAI, explain how and why they use it, and discuss syllabus inclusion and transparency with students.Source: E Educators using GenAI Interview… · p. 38

Learning and equity

• The interview asks whether course learning objectives changed and what positive or negative outcomes and equity issues educators have observed.Source: E Educators using GenAI Interview… · p. 38

Future plans and experiences

• Educators are asked whether they plan to expand GenAI use, what others at their institutions may do, and whether they have additional or memorable course experiences to share.Source: E Educators using GenAI Interview… · p. 38

F Extraction Survey Questions

• The form records inclusion or exclusion decisions, a complete BibTeX entry, the article’s contribution and key results, and additional notes.It offers specific exclusion categories, including too short, K-12, not GenAI, not computing education, no intervention, or no empirical human evidence.Source: F Extraction Survey Questions · pp. 38–39
• Study-context fields cover supervised or unsupervised setting, author affiliation type and country, participant country, education level, number, and description.The form defines supervised studies as conducted in a research lab and unsupervised studies as conducted in a less restricted environment.Source: F Extraction Survey Questions · pp. 38–39
• It asks about the work’s motivation, tools, research questions, programming languages, and whether data are evaluated qualitatively or quantitatively.The form separately asks how authors motivate the work and lists tool and language options.Source: F Extraction Survey Questions · p. 39
• The quality assessment covers whether research questions, process, and results are clearly presented and whether limitations receive an explicit subsection.It distinguishes an explicit limitations subsection from limitations mentioned elsewhere or not mentioned.Source: F Extraction Survey Questions · p. 39
• The form asks how instructors incorporate GenAI and why, and what changes in software-development skill expectations and future computing competencies are discussed.It also requests study limitations and allows additional notes on relevant aspects not captured by other fields.Source: F Extraction Survey Questions · p. 39