Balancing the Risks and Rewards of AI Integration for Michigan Teachers

Introduction

Artificial Intelligence (AI) has emerged as a transformative force in various sectors, and education is no exception. With the integration of AI-driven tools and technologies into classrooms, the educational landscape is undergoing a significant paradigm shift. Educators, as pivotal stakeholders in this evolution, play a crucial role in shaping the future of education through their perceptions and utilization of AI. This study seeks to delve into the intricate world of educators’ perceptions and usage habits with AI in educational settings.

AI applications in education encompass a broad spectrum, from intelligent tutoring systems and automated grading tools to personalized learning platforms and data analytics for student performance assessment (Chen, Chen, and Lin, 2020). These innovations have the potential to enhance teaching and learning experiences, optimize administrative processes, and promote equitable access to education (Baidoo-Anu & Ansah, 2023). However, their effectiveness and impact depend significantly on how educators perceive AI’s role in education and the extent to which they incorporate AI technologies into their teaching practices (Alam, 2021; Celik, 2023; Fitiria, 2021).

Understanding the viewpoints of educators regarding AI in education is essential as it can inform the design of AI solutions that cater to their needs and preferences. Furthermore, it sheds light on the barriers and challenges educators face when adopting AI technologies, such as concerns about job displacement, privacy issues, and areas for professional development.

Significance of the Study

The significance of this study resides in its potential to unveil the intricate interplay between educators and artificial intelligence (AI), a dynamic that stands poised to reshape the educational landscape. This research is particularly important due to the recognition that educators’ perspectives are central, as their acceptance and integration of AI tools will inevitably influence the integration of these technologies into educational settings.

Foremost, this investigation promises to enrich the ongoing scholarly discourse surrounding AI’s role in education by providing valuable insights into educators’ viewpoints. Through an exploration of their usage habits and attitudes, this study seeks to uncover whether educators perceive AI as a valuable complement to their teaching methodologies or as a potential disruptor of their profession. These nuanced understandings can serve as guiding beacons for policymakers, curriculum developers, and technology providers, enabling them to tailor AI solutions that harmonize with the diverse needs and preferences of educators in various educational contexts.

Moreover, the findings of this study carry practical implications, particularly in the domain of teacher training and professional development programs. If educators express reservations or misconceptions regarding AI, this research can pave the way for targeted training initiatives aimed at bridging knowledge gaps and instilling confidence in the effective use of AI tools. Conversely, if educators exhibit enthusiasm for AI, training programs can harness this enthusiasm to enhance the quality and impact of their teaching practices.

On a broader scale, the outcomes of this study can provide valuable guidance to educational institutions regarding the optimal integration of AI solutions. This guidance can address concerns related to data privacy, equity, and the effectiveness of AI-driven pedagogical approaches. Ultimately, the overarching objective is to facilitate the seamless integration of AI into education, ensuring that it serves as a positive transformative force benefiting both educators and learners alike.

Although the field of research on educators’ perceptions of AI in education is still emerging, several studies have initiated the exploration of potential trends and challenges. 

A 2021 meta-analysis conducted by Zheng, Niu, Zhong, & Gyasi indicated that students can derive significant academic benefits from AI-powered tools, including substantial gains in achievement and modest improvements in learning perceptions. Additionally, Kumar & Raman (2022) found that students in higher education harbor positive perceptions about using AI technology for learning in the classroom but believe it should not be employed in high-stakes evaluations, admissions to college, or examinations.

Gocen’s (2020)  phenomenological work looking at perceptions of AI amongst different job sectors had interesting findings; it revealed predominantly positive perceptions toward AI among participants when looking at participants over all job sectors, but found that teachers and those in academia expressed concerns about the future of education and the impacts this technology might have. These findings were in line with Cukarova, Luckin, and Kent’s 2020 research, which also highlighted teachers’ reservations regarding AI-assisted learning and the usage of these tools for various purposes.

Lin (2022) and Sadiku et al. (2021) identified effective teaching designs for AI and discussed barriers and facilitators to participation, interactive design thinking processes, and the alignment of AI knowledge with social good. They also underscored AI’s potential to be a transformative force in the educational design sphere.

Kshirsagar (2022) delved into the potential of AI to enhance teaching and learning, addressing its integration within educational institutions and its impact on students’ adoption. Findings from this study included the development of a model based on the implementation of AI tools that indicated a hybrid approach of learning with AI evaluative tools and content generation with instructor supervision and facilitation would likely be an extremely effective practice.

Xue and Wang (2022) found that teachers recognize AI’s role in reducing teaching workloads and enhancing information literacy, with the majority considering it a valuable tool for their professional development. Their research also uncovered significant relationships between technology usage, familiarity, and dispositions toward technology, with teachers who used and were more familiar with technology generally holding more positive opinions about AI and technology knowledge in general.

Most recently, Celik (2023) conducted a study on teachers’ ability to integrate AI tools into their teaching and curriculum development. The findings suggested that teachers with deeper familiarity with AI-based tools gain a clearer grasp of how AI can contribute to teaching methods. Moreover, having technological expertise enables teachers to make more informed judgments regarding AI applications. However, possessing technological knowledge alone isn’t adequate for effectively integrating AI tools into education. To use AI efficiently in teaching, technological knowledge becomes more impactful when paired with pedagogical understanding, as demonstrated in technological pedagogical knowledge.

In conclusion, educators’ perceptions of AI in education tend to lean slightly toward the positive across all age and grade levels, and some research indicates benefits for students using this technology in the classroom. Nonetheless, it is evident that there are significant concerns and challenges that require careful consideration and resolution.

Research Questions

Two large overarching research questions guided this study and were addressed using mixed methods, which are described in the next section. 

1. What relationships exist between educators’ job types, usage habits, and opinions regarding Artificial Intelligence?
2. What were educators’ perceptions regarding AI in education?

These research questions are largely exploratory, enabling other researchers to build upon the findings using more precise investigations and robust designs.

Methodology 

Over the course of 6 months, Michigan Virtual collected data from educators through an anonymous survey and scraped data from professional learning courses regarding Artificial Intelligence. These data were both quantitative and qualitative, making this a mixed methods study.

The two data sources and the types of analyses associated with each are shown below in Table 1.

Data Source	N	Collection Period	Analytics	Data Type(s)
Survey	125	One time over six months	Descriptive Statistics Pearson R Correlations ANOVAs T-Tests Thematic Analysis	Quantitative Qualitative
Course Discussion Boards	52	Cumulative total over 4 months	Thematic Analysis Descriptive Statistics	Quantitative Qualitative

In addition to using qualitative thematic analysis and coding, multivariate statistical approaches, and mixed-methods holistic synthesis, this study also experimented with using AI to help analyze data where applicable. These AI-powered analyses are described alongside the researcher’s analyses in their own clearly defined sections; it is important to note that AI was NOT used to analyze any of the raw data for the findings, but rather it was used to synthesize what the researcher had already analyzed to provide a summary of the findings.

Results

Below are the results of this study, separated by data source, followed by a synthesis section that summarizes the findings holistically. It is important to note that all data sources are independent of each other, and there is no overlap in participants per data source.

As noted before, the synthesis section is divided into two subsections: one that is the researcher’s synthesis of the findings and one that was made with Artificial Intelligence tools (ChatGPT, Notion AI, and Google Bard). 

Survey Results 

The survey captured data from 125 participants that included their professions, AI usage habits, and opinions on AI. It was distributed by Michigan Virtual via social media and email, taking advantage of its network of educators and innovators in the education space. Responses were captured over a period of about six months in the spring of 2023.

Demographics

The largest group of participants were K-12 teachers (34), followed closely by education administrators (33), “other” education positions (30), and then curriculum designers (17); the other demographic groups comprised approximately 20% of the remaining frequencies in the sample combined.  Figure 1 below is a visual representation of the data.

This distribution of job types was likely due to the network that Michigan Virtual has access to and the convenience of the sampling method that was used. 

Given additional time or reach, a more equal distribution may have been reached; however, given the timeframe, resources, and exploratory goals of the study, this sample was sufficient in representing Michigan Virtual’s direct network of educators.

Artificial Intelligence Usage

Respondents were questioned about their usage habits of one of the most popular AI tools, ChatGPT, as well as general AI usage, in addition to their general disposition about the technology. These results are discussed below.

Overall, 58% of respondents indicated that they had used ChatGPT, and 38% of participants indicated they had used AI in the classroom or in other aspects of their jobs. 

ANOVA tests revealed that there was no significant difference in ChatGPT usage between job categories, but there were differences in general AI usage in the classroom or on the job between job categories (p < .05, F = 2.230). 

While there were some slight differences in usage between most of the groups, the largest differences between groups were between K-12 teachers and the other groups or between the two groups of college educators and education researchers and all the other groups.  

Below is a line graph (Figure 2) of the two different types of AI use, both general AI use and ChatGPT usage specifically.

These initial findings in the ANOVA regarding group differences led the researcher to further examine the groups because of low n in a few of the job categories and the differences between them, specifically because the groups with low n scored very high, and the other groups besides K-12 teachers scored similarly to each other. 

Even though ANOVA is robust enough (able to avoid error) to account for these low n and no data normality assumptions were violated, low n in groups can still be impactful; therefore, reorganizing the data in this way allowed for more observable differences and relationships between groups. 

The easiest way to look at differences was to categorize the respondents into two categories: K-12 teachers and non-K-12 educators, which is shown below in Figure 3.

When job categories were sorted into a binary form (non-K-12 educators and K-12 educators), these differences in usage habits (general AI and ChatGPT) were present on both dependent variables and to a greater degree.

Table 2 (below) shows the significant mean differences between the two groups regarding the two types of usage. K-12 teachers used both AI (F = 1.697  , t = 1.996 , p <.05 ) and ChatGPT (F = 43.88  , t =3.338 , p <001) much less than non K-12 educators.

	Educator Type	N	Percentage of Use	Std. Deviation	Std. Error Mean
ChatGPT_Use	Non-K-12	91	64%	.483	.051
	K-12 Teacher	34	44%	.504	.086
AI_Use	Non-K-12	89	53%	.502	.053
	K-12 Teacher	34	21%	.410	.070

In addition to the group differences, a Pearson R correlation was conducted to determine the strength of the relationship between the variables (Educatory Type and AI Use/ChatGPT Use), resulting in three significant correlations:

1. ChatGPT Use and AI Use are strongly correlated (R =.658, p < .001)
2. ChatGPT Use and Educator Type are very mildly correlated (R =-.177, p = .048)
3. AI Use and Educator Type have a moderately low correlation (R =-.298, p < .001)

This means a direct relationship exists between general AI usage and Chat GPT usage, indicating that if you use one, you are also likely to use the other. Furthermore, the job type of a K-12 teacher may play a small role in whether an educator makes use of either ChatGPT or AI tools due to the small but significant relationships present between the two variables.

Overall findings in this area indicate that K-12 teachers reported the lowest use in both areas (ChatGPT and General AI usage), while education researchers and college educators reported the highest use in both areas. Almost all job types used ChatGPT more than other general AI tools, though the use of both was highly related to each other. When examining relationships and differences between groups on the binary classification of K-12 Educator vs. non-K-12 Educator, differences became more pronounced, with larger differences between the groups indicating that group membership may play a role in the usage of tools.

AI Disposition

The overall disposition of participants regarding AI was collected on a 5-point scale, with 1 being very negative and 5 being very positive. Below are distributions of the overall disposition of participants regarding AI by job type in categorized and aggregated formats, respectively.

The distribution of scores in Figure 4 show that K-12 Educators had more respondents in the two negative categories and the only group with respondents answering “Very Negative”.

The Education Administrators and Other categories had the most positively skewed responses.

Below in Figure 5, this same distribution, along with a curve, helps easily show the differences and similarities between the groups.

Figure 6 shows the average score across all of the job categories side-by-side.

The average disposition across all job categories was 3.41, or slightly above neutral, with a standard deviation of 1.04.

A univariate ANOVA revealed that there were significant differences in AI disposition between job groups (p < .001, F = 8.001). No statistical assumptions of homogeneity were violated, even though some job groups were of small n. 

Overall, on a 5-point scale, Administrators had the highest average outlook on AI (3.88), which was slightly above neutral and leaning toward favorable. K-12 Teachers had the lowest average outlook (2.53), which was below neutral and leaning heavily towards negative dispositions (Figure 6). The difference between the highest (Administrators) and the lowest (K-12 Teachers) groups was highly significant (p <.001). 

Most of the other significant differences in the model were between the K-12 Teachers and other groups because of the lower score from the K-12 Teacher group. Curriculum Designers, Administrators, and Other Education-related positions all reported similar average scores of slightly above neutral.

A table of these average disposition scores is shown below in Table 3.

Dependent Variable:   Outlook_AI (1 = Very Negative, 5 = Very Positive)			95% Confidence Interval	95% Confidence Interval
Job_Category	Mean	Std. Error	Lower Bound	Upper Bound
College educator	3.000	.640	1.732	4.268
Curriculum Designer	3.765	.220	3.330	4.200
Education Administration	3.879	.158	3.567	4.191
Education Researcher	3.000	.906	1.207	4.793
Instructional Coordinator	3.375	.320	2.741	4.009
Other Education-related position	3.733	.165	3.406	4.061
Teacher (K-12)	2.529	.155	2.222	2.837

As with the previous analysis regarding AI usage, the researcher determined that it would be useful to bin participants by binary job types (K-12 and non-K-12 educators) for further analysis. Once binned, the overall difference in disposition between the groups measured by t-test indicated that non-K-12 educators (3.74) had more favorable views on AI as opposed to K-12 educators (2.53) (F = 2.901, t = 6.691, p <.001 ).

	Teacher_Binary	N	Mean	Std. Deviation	Std. Error Mean
Outlook_AI	Non-K-12	91	3.74	.841	.088
	K-12 Teacher	34	2.53	1.051	.180

Pearson R correlations on the binary job types and AI dispositions were also conducted in order to determine the strength of the relationship between group membership (K-12 or non-K-12 educator) and AI disposition, resulting in a moderate (r =-.514), significant (p <.001) inverse relationship between the two variables. 

Overall, these analyses indicate that K-12 teachers have significantly lower dispositions regarding AI than their non K-12 counterparts and that this relationship is moderately strong.

Overall Relationships

Pearson R correlations were also conducted on usage and disposition data; ChatGPT usage had a significant (p <.001)  r of .361 with overall disposition, and AI Usage had a significant r (p<.001) of .440 with overall disposition; while these are significant, they are on the moderately low side of strength which means that while they may be related, they aren’t exceptionally strong relationships or a single best “predictor” of the other. This means that, for both ChatGPT usage and AI Usage in general, if the frequency of use increases, the participants’ general disposition about AI increases as well (and vice versa), with general AI Usage having the stronger relationship. 

Below (Figure 7), is a relationship map that graphically represents the relationships present in this dataset, and it sorts participants into K-12 teachers and non K-12 educators as the primary grouping variable.

The figure above shows how these groups responded to the itemized questions, and uses lines of increasing thickness to show stronger relationships between variables. It shows relationships between teacher types, usage, and disposition. 

The variables are listed in the top right of the figure by color; each sphere of that color represents the categorical responses in its respective variable. For instance, the blue “No” circle represents the categorical response of “No” in AI usage, while the purple “Slightly Positive” circle represents the categorical response of “Slightly Positive” in the AI Outlook variable.

The strongest relationships are as follows:

1. Non-K-12 educators answering yes to both usage items (general AI use and Chat GPT Use), and having neutral, slightly positive, or very positive responses. No non-K-12 educators were in the very negative disposition responses.
2. K-12 educators answering no to both usage items (general AI use and Chat GPT Use), and having neutral, slightly negative, or very negative responses regarding disposition on AI.

Opinions Regarding ChatGPT and AI

Participants were asked the following short answer survey questions in addition to the quantitative questions previously discussed above:

1. What is your opinion regarding ChatGPT, specifically with its usage in education by teachers and other education professionals?
2. What is your opinion about ChatGPT and its relationship with education, specifically with teachers and students?
3. Please describe how you have used AI in your job (if applicable).
4. What is your opinion regarding the general use of AI and its relationship with education?

The answers to these questions were analyzed and will be discussed below both by question, and then holistically. The analytical method categorized responses based on overall theme, meaning that a single response would only fit into a single code category. The top three categories by frequencies of responses are included.

Question 1: What is your opinion regarding ChatGPT, specifically with its usage in education by teachers and other education professionals? Only those that responded “Yes” to ChatGPT usage were included in these responses (N=73)

Code Category	Frequency	Predominant Groups	Description
Potential Benefits for Students and Learning	38%  (N=27)	Administrators Curriculum Designers Other Ed. Professionals	Responses that suggest ChatGPT has the potential to provide a new and innovative way to engage students, provide personalized learning experiences, and support teachers in delivering instruction. These respondents believe that ChatGPT has potential to help students who struggle with traditional instruction methods, and that it can offer more efficient and effective ways to learn.
Concerns and Risks	30%  (N= 22)	K-12 Teachers	Responses that express concerns and risks associated with the use of ChatGPT in education. These concerns include reliability and accuracy of ChatGPT’s responses, potential for ChatGPT to replace human teachers or exacerbate existing inequalities in education, and ethical concerns around the use of artificial intelligence in education.
Support for Teacher Roles	17%  (N=12)	Administrators Instructional Coordinators K-12 Teachers	Responses that suggest ChatGPT can be used to support the work of teachers rather than replace them. Respondents indicated that ChatGPT can provide additional resources and assistance in lesson planning and delivery, and can help to reduce teacher workload by automating some tasks. Some also believe that ChatGPT can help to address teacher shortages by providing support for existing teachers.

Question 2: What is your opinion about ChatGPT and its relationship with education, specifically with teachers and students?  Only those that responded “Yes” to ChatGPT usage were included in these responses (N=73)

Code Category	Frequency	Predominant Groups	Description
Positive Views	45%  (N=33)	Administrators Other Ed. Professionals	Responses that expressed positive views about the potential of ChatGPT in enhancing education for both teachers and students. These responses highlighted the benefits of ChatGPT in facilitating communication, providing access to information, and improving learning outcomes.
Concerns about Reliability and Accuracy	30%  (N=22)	K-12 Educators	Responses that expressed concerns about the reliability and accuracy of ChatGPT in providing educational content. These responses highlighted the potential for bias and errors in the responses generated by ChatGPT, which could negatively impact learning outcomes. Some educators were particularly passionate about this.
Skepticism about Usefulness in Education	15%  (N=11)	K-12 Educators Curriculum Designers Other Ed. Professionals	Responses that expressed skepticism about the usefulness of ChatGPT in education. These responses highlighted the limitations of ChatGPT in providing personalized and interactive learning experiences for students, and the potential for ChatGPT to replace human teachers

Question 3: Please describe how you have used AI in your job?  Only those that responded “Yes” to General AI usage were included in these responses (N=48)

Code Category	Frequency	Predominant Groups	Description
Writing Tasks	40% (N = 19)	Curriculum Designers Administrators Other Ed. Professionals	Responses included things such as drafting emails, creating social media posts, and generating prose for personal writing.
Curriculum Development	25%  (N =12)	Curriculum Designers Administrators	Responses included such things as lesson planning, creating assessments, and developing outlines and discussion prompts.
Proofreading and Editing	20% (N =10)	—–	Responses included reviewing writing for grammar and style, checking for academic honesty, and using AI as a, “second pair of eyes.”

Question 4: What is your opinion regarding the general use of AI and its relationship with education? Only those that responded “Yes” to General AI usage were included in these responses (N=48)

Code Category	Frequency	Predominant Groups	Description
Neutral/Cautious Views	40%  (N =19)	—–	These opinions express a wait-and-see approach to AI in education. They acknowledge the potential of AI but also express concerns about its negative effects or uncertainty about how it will be used. Neutrality was implicitly or explicitly mentioned.
Positive Views	30% (N =14)	Curriculum Designers Other Ed. Professionals Administrators	These are opinions that view AI as a useful tool that can enhance education, as long as it is used appropriately and responsibly.
Mixed Views	20% (N =10)	College Educators K-12 Educators Other Ed. Professionals	These opinions acknowledge both the positive and negative potential of AI in education and stress the importance of responsible use and clear communication among stakeholders. This group was differently coded than Cautious/Neutral because it specifically indicated that they saw both good and bad things about AI, but did not specifically state neutrality.
Negative Views	10% (N =5)	K-12 Educators Administrators Instructional Coordinators	These opinions view AI in education as a negative influence that discourages critical thinking and undermines the role of educators.

Survey Data Synthesis

Nearly 2/3rds of the participants have used ChatGPT and around ⅓ have used AI in aspects of their jobs. Across almost all groups, more participants made use of ChatGPT than other general AI tools. K-12 educators seem to display the most hesitancy and negativity around AI usage, while administrators and curriculum designers, and other education professionals have more positive dispositions around the technology, while also using it more. 

Those who use AI technologies (both ChatGPT or other AI tools) also tend to have stronger positive attitudes towards it, both from a qualitative and quantitative standpoint, which strengthens the findings from the moderate correlation that was found between usage and general disposition around AI, which indicates that increased familiarity is related to reduced negativity towards the tech. In contrast, those educators that use AI tools (including, but not limited to ChatGPT) less frequently tend to have more negative attitudes about the tech, again, supported by correlational and qualitative data. K-12 Teachers have the least usage of AI tools and Chat GPT, and have significantly less favorable attitudes towards AI as well.

While these relationships do not infer that familiarity causes better attitudes, it does imply that a there is a clear and measurable relationship between the variables.

Primary use cases of AI for participants are predominately reserved for short writing tasks such as emails or social media posts, various components of communication, and information synthesis, which is in line with what many participants described as benefits to using tech such as ChatGPT. Secondary use cases included using AI tools for content generation and curriculum design and development. These groups tended to be outside the K-12 educator group.

Opinions around AI technology were mixed overall amongst those who used the tools, with most participants acknowledging the potential benefits and hazards of adopting these technologies in education.

A summary table of the findings is presented below.

Group	AI Adoption	AI Opinion	Dominant Views	AI Use Cases
K-12 Teachers	Low	Unfavorable	Students will use it as a crutch Responsible and ethical use is paramount	Limited Use Curriculum Development Proofreading and Editing
Administrators	Moderate	Favorable	Useful tool for enhancing teaching and learning Clear policy on use cases will address concerns	High Use Curriculum Development Writing Tasks
Other Education Professionals	Moderately High	Favorable	Useful in education and other areas of work Responsible use and “wait and see” attitudes on complete integration of any and all tech	High Use Writing Tasks
Curriculum Designers	Moderate	Slightly Favorable	Very useful for generating content for education	Moderate Use Curriculum Development Writing Tasks
College Educators	Moderate	Neutral	Mixed views, mostly seeing both potential benefits and downsides to mainstream AI adoption	High Use Curriculum Development
Education Researchers	Moderate	Neutral	Useful tools for variety of purposes	Moderate Use Curriculum Development Writing Tasks Proofreading and Editing
Instructional Coordinators	Moderate	Slightly Favorable	Sees practical use cases for educators, but fear student usage.	Moderate Use Writing Tasks Proofreading and Editing

Discussion Board Results

Demographics 

The participants in this course were educators of all types, K-12 educators and administrators, curriculum developers and designers, education technologists, and educational researchers, though the majority of participants were K-12 educators. 

Data Source Description

Michigan Virtual hosts a professional learning course titled ChatGPT for Educators: An Introduction, which serves as a beginner course for understanding what AI is and what its relationship is with education thus far. As part of their learning objectives, participants are asked to post in a discussion forum about their opinions regarding ChatGPT and AI at the conclusion of the course. 

Seventy-Eight discussion posts, with all their thread replies, were included in the thematic analysis in response to the following:

Reflect and Connect Prompt: This discussion thread provides an opportunity for you to reflect on your learning and connect with other learners in this course. Having gotten some knowledge and experience with ChatGPT, what are you currently feeling? (excitement, fear, curiosity, etc.) What sorts of next steps would you like to take in using (or not using) the tool professionally? 

Five major themes were extracted from the data in the discussion posts, and they are discussed below in a summary table for convenience of data presentation.

Theme	Number of Coded Responses	Key Points
Excitement and Potential	31 (39%)	– Excitement about ChatGPT’s potential in education. – Perception of revolutionizing teaching and learning. – Enhancement of critical thinking skills and personalized learning. – Interest in mastering ChatGPT for effective usage.
Concerns	24 (31%)	– Apprehensions about student misuse, plagiarism, critical thinking decline. – Reflection on AI’s societal impact. – Emphasis on independent problem-solving skills.
Ethics and Responsibility	10 (13%)	– Exploration of ethical use of ChatGPT. – Advocacy for responsible AI use education. – Awareness of potential misuse. – Focus on teaching ethical AI utilization.
Applicability and Limitations	7 (9.5%)	– Analysis of ChatGPT’s practicality across educational settings. – Recognition of limitations, especially in complex subjects. – Importance of verifying ChatGPT’s information. – Applicability for basic concepts.
Changes in Teaching and Learning	6 (7.5%)	– Deliberation on ChatGPT’s transformative potential. – Encouragement to embrace technology in education. – Need for updated policies and guidelines. – Emphasis on moving towards student-centered learning.

Excitement and Potential (31 posts)

These posts composed the largest group, and expressed excitement about the potential of ChatGPT in education. Participants believe that ChatGPT can help them and their students and that it can change how teaching and learning happen. They are interested in exploring the tool and learning how to use it effectively.

“ChatGPT has the potential to revolutionize the way we teach and learn. It can help students develop critical thinking skills, and it can make learning more engaging and fun.”

“ChatGPT can help us be more efficient, more creative, and more effective in teaching and learning. It has the potential to make learning more student-centered and personalized.”

Concerns (24 posts)

The second largest group of posts expressed concerns about using ChatGPT in education. They worry about the potential for student misuse, plagiarism, the elimination of critical thinking, truth, and the impact of AI on society.

“I think ChatGPT could be helpful, but I also worry about the impact on critical thinking and the potential for plagiarism.”

“There’s a risk that students could become too reliant on ChatGPT and stop thinking for themselves. We need to make sure that students are still developing their own problem-solving skills.”

Ethics and Responsibility (10 posts)

These participants discussed the ethics of using ChatGPT and the responsibility that comes with it. They are concerned about the potential for misuse and the importance of teaching critical thinking and responsible use of the tool.

“We need to be mindful of the potential for misuse of ChatGPT. It’s important to teach students how to use it responsibly and ethically.”

Applicability and Limitations (7 posts)

Posts in this category discussed the applicability of ChatGPT in different settings, such as middle school and elementary, and the limitations of the tool in explaining mathematics to someone who doesn’t already know the topic, or how to teach art with it. They also mentioned the need to verify the accuracy of the information provided by ChatGPT.

“ChatGPT is a great tool for teaching basic concepts, but it might not be as useful for more complex topics that require a deeper understanding.”

Changes in Teaching and Learning (6 posts)

This last, smallest category included participants discussing how ChatGPT could change teaching and learning and how educators need to embrace new technology instead of running from it; there was no clear positive or negative tone, and thus why they did not fit into the two largest categories. They did, however, all mention the need for new policies and guidelines for responsible use of ChatGPT.

“ChatGPT could help us move away from a traditional lecture-based model of teaching and toward more student-centered learning, but we have to be careful how we move forward with this.”

Course Discussion Boards Synthesis

Overall, the discussion posts on ChatGPT in education revealed a mix of excitement, potential, and concerns about the tool. Around half of the pool has positive feelings towards the technology and around half have negative or apprehensive views about it.

Most participants are interested in exploring the tool and learning how to use it effectively, but they also recognize the need for ethical use, responsible teaching, and verification of information provided by ChatGPT.

Synthesis

This study collected data from two sources: a course discussion board for educators and a short survey on educators’ perceptions, opinions, and usage habits on Artificial Intelligence. Below are two subsections that will synthesize the findings of both data sources: one is written by the researcher and took approximately two hours, while the other is written by artificial intelligence and took 31 minutes. 

This approach was taken to show readers the capabilities and efficiencies of artificial intelligence, and allow a compare and contrast between the two syntheses. 

Researcher Synthesis 

This section will synthesize and summarize the findings of this study, organized by research question. A discussion of the implications and recommendations from these findings can be found in the following Conclusions section.

Research Question 1: What relationships exist between educators’ job types, usage habits, and opinions regarding Artificial Intelligence?

Findings indicate that there are significant differences in AI usage habits between K-12 classroom educators and those that are not, with K-12 teachers reporting lower usage of general AI tools and ChatGPT. Additionally, there was a moderately strong correlation between being a K-12 teacher and having more negative opinions about the technology (r =-.514), as well as lower usage habits to a smaller degree (ChatGPT Use r = -.177; AI Use r = -.298).

Administrators and other educational professionals had more favorable opinions and spoke more positively about AI, in addition to having made more use of the technology on average. 

The majority of K-12 teachers were quicker to point out more concerns and weaknesses of the tools, in general, and when they did speak about use cases, they spoke more about students using the tools rather than the teachers using it themselves. The minority of teachers that did speak positively about AI were those who had used it more and were more familiar with the technology, and they tended to have a plethora of use-cases that they spoke about for AI in the classroom.

Administrators spoke more positively about AI tools and addressed opportunities for use of it by both teachers and students, and when they spoke about concerns and weaknesses, they generally focused on how teachers need to adapt and change their teaching methods to account for the advent and adoption of AI in the classroom. 

The general relationship between usage habits and perceptions was moderate, but significant, indicating that the more an educator (of any type) uses artificial intelligence technology, the more likely they are to have a more favorable opinion of it and vice versa. 

Research Question 2: What were educators’ perceptions regarding AI in education?

The findings to this research question are discussed by major themes that were present in the data.

Strengths and Opportunities in AI

Both data sources in this study indicate that there are opportunities for the leveraging of AI tools in many aspects of education, many of them overlapping. They can be organized into two broad categories: Personalized Learning and Teacher Job Function.

Personalized Learning 

Every single data source pointed to this use case as the strongest opportunity for AI in the classroom. Whether it was the administrators, curriculum designers, teachers, or other education professionals, if they were talking about using AI in a positive manner, then they most likely were talking about how it can be used to advance personalized learning in some form or fashion.

Those who spoke about this generally had concrete evidence from direct use or observation of AI tools that have been used to help students learn at their own pace, increase their agency in what and how they are learning, or simply provide things like smart tutor for individuals.

Advances in the field of personalized learning were seen as the greatest opportunity for AI in the classroom, but also were generally mentioned alongside a caveat of caution, specifically because participants mentioned that unless safeguards were in place, this type of tech could be abused by students or cause harm to students in a variety of unforeseen ways due to AI’s inability to actually understand humans, their emotions, and cognition in a way that a human teacher can.

Teacher Job Function

The other major theme that was present across both of the data sources was the idea that AI can aid teachers in ways that can revolutionize how they do their jobs: increasing productivity by cutting down lesson planning time in huge amounts, serving as a ‘jumping off point’ for content creation, and even assisting in assessment in a variety of ways.

Participants were quick to point out very specific use cases in which teachers could leverage AI to impact their job function. Whether it was describing how a teacher could use AI to quickly design 10 assignments on math to differentiate to 3 different reading levels, 5 different topics of interest, and two different levels in Bloom’s Taxonomy of higher order thinking or simply having students use AI as an interactive librarian that can serve as a keeper of knowledge, freeing up the teacher to act as a true facilitator of learning that research has been clamoring for teachers to engage in for years.

This strength was seen as slightly less revolutionary than massive jumps in personalized learning, but also safer, as it would be supervised AI use by an adult that can control what is presented before the students rather than on-demand and instant AI interaction with students.

Threats and Weaknesses of AI

The data in this study indicated that there were two overarching groups of concerns and threats regarding the use of AI in education, and they are discussed below.

Ethics and Equity

Every data source in this study referred to plagiarism and/or academic dishonesty as a large concern with tools such as ChatGPT. While some believe that technology like this should be controlled and/or banned, most tended to believe that it was the responsibility of the field of education to step up its game to a point where the use of such tools would not give students a shortcut to success, but rather provide them with tools to solve and address more complex problems and issues; assessment and evaluation of student learning needs to change.

In addition to concerns with cheating, many also spoke about how these tools will continue to evolve and improve, and that if access to these tools is not provided for everyone, then it could imbalance teaching and learning in unforeseeable ways, widening the gap between high and low SES students. Again, when this was spoken of, participants tended to indicate that it would need to be a decision from the top to make sure that these types of things do not happen, and that education must start making decisions about these things now, rather than wait.

Human Experience

The third theme that was present in the data was the idea that AI cannot and will not replicate the human experience of learning with a live teacher, or that AI cannot be trusted to simply complete tasks that interface with humans all on their own with no supervision.

Participants indicated that only a human teacher can have intimate knowledge of their students in the classroom, and that while AI may be a great tool, it can’t simply replace that level of connection that a teacher can create, which is one of the most effective ways of teaching and learning. 

Furthermore, because AI has no true understanding of the human experience, emotions, cognition, or the world we live in, it does not always communicate in effective or safe ways, rather it communicates based on a set of rules and the information that it has access to. For these reasons, participants agreed almost unanimously, across all data sources, that this type of technology can’t replace an expert in their field when it comes to all the functions that expert fulfills– AI might can pass the test, regurgitate information, analyze information, and provide suggestions and answers, but it won’t replace a human’s experience in the application of all those amazing capabilities in their field.

Artificial Intelligence Synthesis

This section was generated in 31 minutes using a combination of ChatGPT, Notion AI, and Google’s Bard (all free versions). It went through 26 prompt revisions, and is represented here in its entirety, unedited by the researcher.

Research Question 1: Relationships between Educators’ Job Types, Usage Habits, and Opinions

The data in this study examined the relationships between educators’ job types, usage habits, and opinions regarding Artificial Intelligence (AI) in education. The findings revealed varying levels of AI usage and disposition among different job categories within the education sector.

K-12 teachers reported the lowest usage of AI tools compared to education researchers, college educators, administrators, and curriculum designers. This difference in usage habits was also reflected in their attitudes towards AI. K-12 teachers exhibited less favorable attitudes towards AI, while administrators and curriculum designers displayed more positive attitudes and utilized AI more frequently in their work.

The data further highlighted that educators who use AI technologies tend to have stronger positive attitudes towards them. Familiarity with AI was found to be related to reduced negativity towards the technology. K-12 teachers, who had the least usage of AI tools, also exhibited less favorable attitudes towards AI. These findings emphasize the importance of considering educators’ job types and usage habits when implementing AI technologies in educational settings.

Research Question 2: Educators’ Perceptions regarding AI in Education

This study explored educators’ perceptions and attitudes towards AI in education. The findings revealed mixed opinions among participants who used AI tools.

On one hand, participants expressed excitement about the potential of AI to revolutionize teaching and learning. They believed that AI could develop critical thinking skills, personalize learning experiences, and free up teachers to focus on evaluating student questions and facilitating learning. They recognized the benefits of AI in enhancing the educational experience for students.

However, concerns were also raised regarding potential challenges associated with AI usage. Participants expressed worries about student misuse, such as cheating and plagiarism. They also highlighted concerns about the elimination of critical thinking, the impact on truth, and the broader societal implications of AI. Ethical considerations were seen as crucial in the use of AI in education, including transparency, accountability, equity, and inclusion.

Participants emphasized the need for responsible use and stressed the importance of establishing policies and guidelines for the ethical and responsible use of AI in educational settings. They recognized that while AI presents opportunities for personalized learning, it must be implemented in a way that safeguards against potential misuse and upholds ethical standards.

In conclusion, this study sheds light on the relationships between educators’ job types, usage habits, and opinions regarding AI in education. It highlights the variation in attitudes towards AI across different job categories and emphasizes the role of familiarity with AI in shaping attitudes. The study also underscores the potential of AI for personalized learning but emphasizes the need for responsible use and ethical considerations. These findings can guide the development and implementation of AI technologies in educational settings, taking into account educators’ perspectives and the importance of ethical considerations.

AI vs Researcher Synthesis 

As a researcher, data scientist, and writer, it was very apparent to me that Artificial Intelligence LLMs like ChatGPT, Bard, Notion, or others, are able to take information from qualitative and quantitative data and synthesize it in a format that is comparable to what expert researchers can produce, and in a considerably shorter amount of time. This is, however, something that requires expertise in scientific communication, knowledge of the desired formatting, and a keen eye for noticing key errors or omissions in the outputs that are produced; there is little chance that I could have produced what is present here with simply one prompt with little to no contextual knowledge or expertise in my field. Context, and iterative prompt revision based on careful output reading and synthesizing are key to producing high-quality content.

This anecdotal evidence would support the idea that LLMs and AI tools have a place in many of the use cases mentioned by participants, but that the more specialized the desired output becomes, the more important the user’s contextual knowledge and expertise becomes.

Conclusions 

This study of educators’ perceptions around AI in education suggests that K-12 teachers are generally more reluctant, concerned, and feel negatively towards AI usage, while administrators and curriculum designers have more positive dispositions towards the technology; this is similar to the findings from Gocen’s (2020) and Cukarova et al. (2020) work in AI.  Furthermore, this study showed that those who use AI technologies in their work tend to have stronger positive attitudes towards it from both a qualitative and quantitative standpoint, suggesting that familiarity decreases negativity towards the tech, which is similar to Xue and Wang’s (2022) findings and almost identical to the findings of Celik’s 2023 robust study.

The study also found that participants believe that AI has the potential to personalize the learning experience for students in ways that can free up teachers to evaluate student questions and work with more detail, becoming more true facilitators of learning, much like the work from Xue and Wang (2022). In addition, educators identified several use cases for AI in the classroom that were similar to the findings of Lin’s study in 2022 and the earlier study by Sadiku et al. in 2021.

However, concerns regarding the equity of the technology in the classroom, the concerns with academic dishonesty, and the need for guidelines and policies on AI in the classroom cannot be ignored, and should be pursued with haste.

Based on these findings, the following recommendations can be made:

1. Provide more opportunities for K-12 teachers to learn about and use AI in their classrooms, in order to reduce hesitancy and increase familiarity with the technology.
2. Encourage the experimentation and use of AI technologies in safe education workspaces to enhance personalized learning and free up teachers to focus on higher-order thinking skills and individualized instruction.
3. Develop safeguards, protocols, and usage standards around AI and its applications in the world to protect against unethical or harmful uses of the technology.
4. Emphasize the importance of responsible use and clear communication among stakeholders when it comes to AI in education.
5. Provide training and resources for educators to effectively and responsibly experiment with and integrate AI technologies into their instruction and curriculum.

These recommendations are similar to those that were provided in Kshirsagar’s recent 2022 study and Celik’s 2023 work, compounding the evidence that steps need to be taken toward adopting changes with regards to AI in education. In addition, these findings resonate with previous Michigan Virtual work on the topic of AI integration, specifically within their Artificial Intelligence Integration Framework for School Districts and its associated Planning Guide. Below is a table including this study’s recommendations with their associated integration component(s) to allow for fidelity of their implementation.

Recommendation	Michigan Virtual AI Framework Component
Provide more opportunities for K-12 teachers to learn about and use AI in their classrooms, in order to reduce hesitancy and increase familiarity with the technology.	Leadership and Vision, Professional Learning
Encourage the use of AI technologies in education to enhance personalized learning and free up teachers to focus on higher-order thinking skills and individualized instruction.	Instructional Framework, Learning Assessments, Business and Technology Operations, and Student Use
Develop safeguards, protocols, and usage standards around AI and its applications in the world to protect against unethical or harmful uses of the technology.	Policy, Ethical, & Legal Considerations
Emphasize the importance of responsible use and clear communication among stakeholders when it comes to AI in education.	Leadership and Vision, Outreach
Provide training and resources for educators to effectively integrate AI technologies into their instruction and curriculum.	Professional Learning

In conclusion, the study suggests that while there are concerns and risks associated with the use of AI in education, the potential benefits for personalized learning and the advancement of 21st-century skills cannot be ignored, as the technology is an incredible tool that will continue to advance and evolve to meet the needs of society. It is important to approach the use of AI in education with caution and responsibility, while also recognizing the potential for positive impact and actively working to incorporate the technology into the classroom. The more we experiment with, integrate, and familiarize ourselves with AI technologies, the better we can utilize them with fidelity, responsibility, and power in the classroom.

References 

Alam, A. (2021, November). Possibilities and apprehensions in the landscape of artificial intelligence in education. In 2021 International Conference on Computational Intelligence and Computing Applications (ICCICA) (pp. 1-8). IEEE.

Baidoo-Anu, D., & Ansah, L. O. (2023). Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and learning. Journal of AI, 7(1), 52-62.

Celik, I. (2023). Towards Intelligent-TPACK: An empirical study on teachers’ professional knowledge to ethically integrate artificial intelligence (AI)-based tools into education. Computers in Human Behavior, 138, 107468.

Chen, L., Chen, P., & Lin, Z. (2020). Artificial intelligence in education: A review. Ieee Access, 8, 75264-75278.

Cukurova, M., Luckin, R., & Kent, C. (2020). Impact of an artificial intelligence research frame on the perceived credibility of educational research evidence. International Journal of Artificial Intelligence in Education, 30, 205-235.

Fitria, T. N. (2021). Artificial Intelligence (AI) In Education: Using AI Tools for Teaching and Learning Process. In Prosiding Seminar Nasional & Call for Paper STIE AAS (Vol. 4, No. 1, pp. 134-147).

Gocen, A., & Aydemir, F. (2020). Artificial Intelligence in Education and Schools. Research on Education and Media, 12, 13 – 21.

Kshirsagar, P.R., Jagannadham, D.B., Alqahtani, H., Noorulhasan Naveed, Q., Islam, S., Thangamani, M., & Dejene, M. (2022). Human Intelligence Analysis through Perception of AI in Teaching and Learning. Computational Intelligence and Neuroscience, 2022.

Kumar, V. R., & Raman, R. (2022). Student Perceptions on Artificial Intelligence (AI) in higher education. In 2022 IEEE Integrated STEM Education Conference (ISEC) (pp. 450-454). IEEE.

Lin, X., Chen, L., Chan, K.K., Peng, S., Chen, X., Xie, S., Liu, J., & Hu, Q. (2022). Teachers’ Perceptions of Teaching Sustainable Artificial Intelligence: A Design Frame Perspective. Sustainability.

National Research Council. 2000. How People Learn: Brain, Mind, Experience, and School: Expanded Edition. Washington, DC: The National Academies Press. https://doi.org/10.17226/9853.

Sadiku, M.N., Ashaolu, T.J., Ajayi-Majebi, A.J., & Musa, S.M. (2021). Artificial Intelligence in Education. International Journal Of Scientific Advances.

Xue, Y., & Wang, Y. (2022). Artificial Intelligence for Education and Teaching. Wireless Communications and Mobile Computing.

Zheng, L., Niu, J., Zhong, L., & Gyasi, J. F. (2021). The effectiveness of artificial intelligence on learning achievement and learning perception: A meta-analysis. Interactive Learning Environments, 1-15.

Appendix

Statistics Tables for AI Usage

			95% Confidence Interval	95% Confidence Interval
Job Type	Mean	Std. Error	Lower Bound	Upper Bound
College educator	1.000	.350	.307	1.693
Curriculum Designer	.588	.120	.351	.826
Education Administration	.667	.086	.496	.837
Education Researcher	1.000	.495	.020	1.980
Instructional Coordinator	.625	.175	.278	.972
Other Education-related position	.600	.090	.421	.779
Teacher (K-12)	.441	.085	.273	.609

			95% Confidence Interval	95% Confidence Interval
Job Type	Mean	Std. Error	Lower Bound	Upper Bound
College educator	1.000	.339	.329	1.671
Curriculum Designer	.647	.116	.417	.877
Education Administration	.500	.085	.332	.668
Education Researcher	1.000	.051	.051	1.949
Instructional Coordinator	.500	.165	.165	.835
Other Education-related position	.448	.272	.272	.624
Teacher (K-12)	.206	.043	.043	.369