Generative AI and Teachers - For Us or Against Us? A Case Study

—We present insightful results of a survey on the adoption of generative artificial intelligence (GenAI) by university teachers in their teaching activities. The transformation of education by GenAI, particularly large language models (LLMs), has been presenting both opportunities and challenges, including cheating by students. We prepared the online survey according to best practices and the questions were created by the authors, who have pedagogy experience. The survey contained 12 questions and a pilot study was first conducted. The survey was then sent to all teachers in multiple departments across different campuses of the university of interest in Sweden: Luleå University of Technology. The survey was available in both Swedish and English. The results show that 35 teachers (more than half) use GenAI out of 67 respondents. Preparation is the teaching activity with the most frequency that GenAI is used for and ChatGPT is the most commonly used GenAI. 59% say it has impacted their teaching, however, 55% say there should be legislation around the use of GenAI, especially as inaccuracies and cheating are the biggest concerns.


I. INTRODUCTION
Recent advances in artificial intelligence (AI), especially generative artificial intelligence (GenAI), have caused a stir in the Education sector around the world [1], [2].ChatGPT 1 , the leading Large Language Model (LLM) by OpenAI, has both been beneficial and controversial.Some of the concerns with GenAI are generating deepfakes and the nature of these "unexplainable" models [11].In spite of these concerns, many recognize the benefits inherent in these technologies [7], [11].In this case study, we seek to understand how university teachers perceive GenAI and investigate the following research questions. 1) To what extent are university teachers open to adopting generative AI in their teaching and classrooms? 2) What is the correlation between the impact of GenAI on teachers' teaching activities and their encouragement of their students to use it?There's increasing study of the impact of GenAI on students [7], [11].It is equally important to study the impact on other stakeholders or teachers' teaching activities in Education [9].United Nations Educational, Scientific and Cultural Organization (UNESCO) hopes GenAI will be a tool that benefits teachers, students, and researchers.Our main contributions include the following: (1) We show through data that university teachers in this case study are open to adopting GenAI; (2) we demonstrate the correlation between the positive impact of GenAI on teachers to their willingness to encourage their students to adopt it; (3) we provide many qualitative examples of comments of teachers on the impact of GenAI, ways they encourage students, and their concerns.
The rest of this paper is organised as follows.The literature review is discussed in Section II.The method employed in this work is described in detail in Section III.The findings are discussed in Section IV.We conclude with closing remarks in Section V.

II. LITERATURE REVIEW
The subject of GenAI in teaching is gaining increasing attention.It's impact on pedagogy cannot be ignored.Recent LLMs, such as ChatGPT, Aurora-M [15], Large Language Model Meta AI (LLaMA)-2 [21] and a host of others, have compelling abilities to generate human-like content, based on their training with big data [11].This has prompted United Nations Educational, Scientific and Cultural Organization (UNESCO) to publish the guidance for GenAI in education and research [11], which builds on their recommendation on the ethics of AI.To gauge the awareness of teachers or educators and their adoption of GenAI, it is useful to conduct a survey, similarly to that done with students [6], [20].
[14] identified some of the best practices for designing the questionnaire for a survey.As simple as it may sound, the ordering of questions is an important consideration [12], [13].These best practices are essential to have quality data from the survey.[14] also show that it is important to design questions to avoid acquiescence bias, which is the endorsement of a statement, regardless of the content.Testing a survey in a pilot study usually improves the quality of the full survey, as emphasized by [16], [17], [4], [5].Using either closed or open questions have their own benefits, as demonstrated by [19], [18], [8].In our work, we combined both types of questions to get the best our of the survey.

III. METHOD
Our chosen method for evaluating the use of GenAI by teachers at LTU is through a survey.It was designed to gather information about their habitual use of GenAI in teaching.In creating the survey, a few factors were taken into consideration to keep the inquiry objective and to avoid phrasings that could potentially skew the answers of the participants, based on best practices [14] .The factors are given below but are not limited to them.
• Keep the questions simple and concise, so as not to produce off-topic answers [10], [3].
• Include a broad, exhaustive list of viewpoints [14] • Avoid questions where people tend to agree or disagree with statements regardless of their actual feelings or beliefs [14] • Avoid single or double negation questions [14].To reach a larger target audience, the survey questionnaire was created in both Swedish and English using an online tool. 2 The Swedish translation was carried out by some of the authors of this work, who are native speakers.The translations in the Results and Appendix sections were machinetranslated and vetted by the native speakers.The survey consists of 12 questions to cover a relatively broad range of concepts that are necessary to answer the research questions.For the purpose of the survey, a simplified definition was provided for GenAI: Generative AI, such as ChatGPT or DALL-E, is a tool that can answer questions and create images and other media based on prompts from the user [11].A copy of the questionnaire is available online. 3  The survey was anonymous and a pilot study involving 8 teachers was initially carried out to ascertain if the forms and questions needed any adjustments.This full study involved sending the online survey via a link to all the teachers in multiple departments in different geographical campuses of LTU.The following are the 12 questions in the survey and their answer options, as created by the authors based on their pedagogy experience and the factors mentioned earlier.Q1 and Q8 were multiple choice questions.For teachers who answer "None" to the first question ("filter question" [14]), they were directed to continue to question 6 onwards.way?And if "No", please say why.

11) Your gender i) Woman ii) Man iii) Non-binary iv) Prefer not to say 12) Your Division and Department
The following LTU departments were involved in filling the survey.

IV. RESULTS AND DISCUSSION
The survey took 3:53 minutes for each teacher to complete on average.From Table I, 32 (48%) of the teachers do not use GenAI in any of their teaching activities.The remaining 35 (52%) use GenAI for one or more teaching activities, where Preparation is the most frequent activity, being 27%.Besides the identified teaching activities, one teacher explained that she uses GenAI "To see if student work is AI generated".The wordcloud of Figure 1 shows that ChatGPT has the lion share of usage with 52% of frequency of mentions (26 out of 50), in answer to Q2.The nearest is DALL-E, with 6%.Gemini, Stable Diffusion Web, and Midjourney come in next at 4% while all the other GenAIs have 2% (only 1 mention).Figure 2 shows the distribution of activities across gender.
Of the 52% who use these tools, there are as many teachers who use them twice or more in a week (i.e. more than once) as there are those who use them less than once a month (31.58%).About 59% say GenAI has impacted their teaching.Indeed, we see a strong positive correlation between the Yes choices of the departments to Q4 (Do you think the use has impacted your teaching?) and Q9 (Will you encourage any of your students to use generative AI?), based on the Spearman's correlation coefficient (ρ) +0.9474, 5 for p = 0.01438 (2-tailed).This implies the association between the two variables can be considered statistically significant.Several examples of the ways teachers feel GenAI has impacted their teaching are given in the appendix.Most (85%) of the 67 teachers do not think AI will replace teachers.Many (55%) say there should be legislation around 5 r = 1 and -1 are perfect positive and negative correlations, respectively the use of GenAI, possibly because of the ethical concerns observed.Inaccuracies (34.36%) and cheating (26.99%) are the two most common concerns teachers have.Figure 3 shows the concerns across the departments.Given that the majority of teachers encourage their students to use GenAI, some of the ways they go about it are listed in the appendix, including these:   preferred not to say.

V. CONCLUSION
We have shown in this case study that teachers are open to adopting GenAI, as over 50% currently use it.We also observe a strong positive correlation between the positive impact of GenAI on their teaching activities and their willingness to encourage their students to adopt GenAI.We agree with the comments of some of the teachers that students "are guaranteed to use it already", therefore we believe teachers should be knowledgeable about these tools in order to provide the appropriate guidance for students.Future work can investigate some of the ways of addressing the concerns of teachers expressed in this study.
counterpart funding from Luleå University of Technology (LTU).

APPENDIX
Examples by teachers of impact on teaching 1) I am able to provide more bang for the bucks and student feedback shows the impact.2) I made my subject knowledge deeper.
3) ChatGPT hjälper mig att sammanfatta innehållet i t ex en workshop eller en föreläsning (ChatGPT helps me summarize the content of, for example, a workshop or a lecture) 4) t.ex.bättre bilder till mitt undervisningsmaterial, hjälp med bra översättning till engelska etc. (for example better pictures for my teaching material, help with a good translation into English, etc.) 5) preparation of slides and text material is more efficient and result more impactful.6) Bättre språk, fler exempel (Better language, more examples) 7) I show its use for students to use properly as a tool, as well as when or how not to use it.8) improved clarity 9) Jag får ett bollplank som kan hjälpa mig som lärare.(I get a sounding board that can help me as a teacher.)10) Jag har använt Canva:s generativ AI för att snabbt ta fram illustrationer till mina powerpoints.Jag tror att det kan ha en viss positiv påverkan för inlärning att få "bildstöd" till anteckningarna.(I have used Canva's generative AI to quickly produce illustrations for my powerpoints.I think having "visual support" for the notes can have some positive impact on learning.)11) Tidsbesparande (Timesaving) 12) Ibland dyker det upp aspekter som jag tidigare inte tänkt på, men som är relevanta.(Sometimes aspects appear that I previously did not think about, but which are relevant.)13) Är ett fantastiskt verktyg att skapa bilder istället för att leta clip-art.Att få hjälp att förklara saker samt som kreativt verktyg i idegenereringsprocessen (Is a great tool to create images instead of looking for clip-art.To get help explaining things and as a creative tool in the idea generation process) 14) Jag har fått en bättre förståelse för hur studenter kan använda det som stöd samt jag har lärt mig att känna igen resultaten i studenters arbete (I have gained a better understanding of how students can use it as support and I have learned to recognize the results of students' work) Examples of other concerns 1) I have had trouble with students copying AI generated information.I am afraid they are not using it as a learning tool, but rather to avoid learning.2) ...students' usage affects the type of tasks I can give them and how I test their knowledge.That's mainly why I use it myself.

1 ) 6 )
Have you used any generative AI in any of your teaching activities (e.g.preparation, teaching, assessment, or none)?i) Preparation ii) Teaching iii) Assessment iv) Research 4 v) Administration vi) None vii) Other 2) Which ones?e.g GenAI, Ex.ChatGPT, DALL-E, Bing AI, Google Bard etc.Others 3) How often do you use the one you use most?i) Once a month ii) Once a week iii) Twice or more a week iv) Less than once a month 4) Do you think the use has impacted your teaching?i) Yes ii) No iii) Not sure 5) Briefly describe the impact on your teaching.Do you think AI will replace teachers in your subject if the trend of AI development continues?i) Yes ii) No iii) Not sure 7) Do you think there should be legislation around the use of generative AI? i) Yes ii) No iii) Ambivalent 8) What are some of your ethical concerns about generative AI? i) Gender bias ii) Racial bias iii) Inaccuracies iv) Cheating v) None vi) Other concerns 9) Will you encourage any of your students to use generative AI (in an ethical manner)?i) Yes ii) No 10) If you answered "Yes" in the previous question, In what 2 Miscrosoft Forms 3 forms.office.com/e/jvrmPPaJJh 4 Undersökning in Swedish, though initially translated as Undervisning.

•
"...I do 1 or 2 sessions on how and why they should use it.I also show them when it can give wrong results and how to fact check it..." • "seed texts, help in checking texts and results; use AI as study buddy" • "Det är ett mycket effektivt hjälpmedel och bör uppmuntras.Är man orolig för fusk så examinerar man studenterna på fel sätt.(It is a very effective aid and should be encouraged.If you are worried about cheating, you are examining the students in the wrong way.)"Overall,37.31%, 55.22%, and 7.46% of the teachers who completed the survey were women, men and those who

Fig. 2 .
Fig. 2. GenAI usage activities across gender out of a total percentage of 100%.

TABLE I RESULTS
IN PERCENTAGE (%).THE GENDER AND DEPARTMENT SECTIONS EACH ADD UP TO THE TOTAL.
AI kan vara mycket hjälpfullt om man har ett koncept/begrepp som man inte förstår men vill ha förklarat så att man sen, självständigt, kan använda konceptet/begreppet. (AI can be very helpful if you have a concept that you don't understand but want explained so that you can then, independently, use the concept.)14) Utmärkt för sammanfattningar och scanning av stora litteraturmängder (Excellent for summaries and scanning large volumes of literature) 15) utkast till texter, utkast till musik, tex.(drafts of texts, drafts of music, e.g.) 16) Det är ett ypperligt bollplank, framförallt då man inte har någon fysisk person att diskutera med, men det kan även ge förstärkning om man använder den då man är studerar i grupp.(It is an excellent sounding board, especially when you have no physical person to discuss with, but it can also provide reinforcement if you use it when you are studying in a group.)17) I produktutveckling, som ett stöd och verktyg.(In product development, as a support and tool.)18) skriv hjälp, hjälp att komma igång med ett arbete, skapa bilder för presentationer, hjälp mot skrivkramp (write help, help to get started with a work, create images for presentations, help against writing cramp) 19) 1. Använda det för att stava rätt i inlämningar.2. För att ha någon att bolla idéer kring studentarbete.3. Fråga om grundläggande koncept som AI kan behärska och förklara.(1.Use it to spell correctly in submissions.2. To have someone to bounce ideas off of student work.3. Ask about basic concepts that AI can master and explain.)20) Vi lärare behöver skapa och formulera premisser som är rimliga för detta, nu när man inte kan backa bandet med AI.Hur -vet jag inte än.(We teachers need to create and formulate premises that are reasonable for this, now that you cannot reverse the trend with AI.How -I don't know yet.)