Thoughtful and Ethical Approaches to Using Generative AI in Schools

Framework

Artificial Intelligence (AI) technologies, particularly conversational agents like ChatGPT, are making significant inroads into various sectors, including education (Iqbal, Ahmed, & Azhar, 2022; Kasnesci et al., 2023; Lo, 2023; Tlili et al., 2023). As these technologies become increasingly sophisticated, understanding their impact and role in educational settings becomes imperative. Teachers and students are among the primary users of these educational technologies (Walton Family Foundation, 2023). However, there is a gap in existing research concerning how these groups perceive and utilize generative AI tools like ChatGPT, particularly in the context of their schooling.

Given the rapid adoption of AI in education, there's an urgent need for research that focuses not just on the technological capabilities but also on the ethical and responsible use of these tools (Akgun & Greenhow, 2021; Mhlanga, 2023). This is especially important considering the range of potential applications and implications for academic integrity, personalized learning, and data privacy (Blose, 2023; Eke, 2023). Furthermore, as K-12 classrooms consider the integration of such technologies into the curriculum, the perspectives of teachers and students—who are the end-users—become invaluable for policy formulation and pedagogical strategies.

Methods

This study addressed three primary research questions: 1) How do teachers and students perceive the effectiveness of tools like ChatGPT in aiding their learning process? 2) For what educational purposes are teachers and students most commonly using generative AI? and 3) What do teachers and students consider to be responsible and ethical use of generative AI in educational settings?

To answer these questions, we conducted focus groups with 15 middle and high school teachers and 15 middle and high school students* from across the U.S. We ensured that the sample was diverse in regard to race/ethnicity and locale (urban vs. rural). We harnessed our partner networks to recruit participants for this study.

In these focus group sessions, we collected a mixture of qualitative and quantitative data to gain a comprehensive understanding of teacher and student perspectives on generative AI. Qualitative data was primarily gathered through open-ended questions and discussions, focusing on topics such as teacher and student attitudes towards generative AI, their perceived utility of the technology in educational settings, and their opinions on ethical and responsible use. This included verbal expressions of comfort or discomfort with the technology, specific examples of how they have used or would like to use generative AI, and their thoughts on what guidelines should govern its use in schools. Additionally, we employed Zoom polls during the sessions to collect quantitative data on certain variables, such as frequency of use, the types of tasks it is used for, and the level of satisfaction or dissatisfaction with generative AI. This multimodal data collection approach aimed to capture the complex array of factors that influence students' experiences and perceptions of generative AI.

For the analysis of data collected during focus groups, we employed Rapid Qualitative Inquiry (RQI), a method particularly well-suited for quickly and iteratively understanding complex issues (Vindrola-Padros, 2021). RQI involves a flexible, team-based approach that allows for immediate data analysis and interpretation, making it possible to adapt the focus group questions or structure in real-time if necessary. Transcripts from the recorded Zoom sessions were analyzed by the Research team who collaboratively coded the data to identify recurring themes and patterns.

The RQI method helped us transform these insights into actionable recommendations and guidelines on a shorter timeline than other qualitative analysis strategies. The use of RQI is particularly beneficial for studying tools like ChatGPT, especially because artificial intelligence and its role in education are rapidly changing.

In addition to the qualitative data from the focus groups, we also analyzed the quantitative data collected through Zoom polls. Descriptive statistics used to summarized the frequency of generative AI use, types of tasks it is used for, and the level of satisfaction among participants. This multi-method analytical approach will provide a comprehensive view, allowing us to offer a more robust interpretation of the data.

*Note that we are currently in process of collecting data for student focus groups. We completed teacher focus groups and analyzed the data.

Results

Key Findings

How do teachers perceive the effectiveness of tools like ChatGPT in aiding their learning process?

Teachers highlighted that AI tools, especially ChatGPT, are becoming integral to contemporary classrooms, offering novel educational solutions. Many found ChatGPT invaluable for tasks ranging from lesson planning to creating math problems, crafting writing prompts, tailoring reading tasks based on student abilities, and assisting in research and language translation. However, despite its versatility, there are concerns. Issues surrounding student engagement and privacy have led some districts to restrict platform access. Furthermore, while ChatGPT's potential in grading was acknowledged for its consistency, it also sparked debate, with some teachers questioning the ethics of AI in assessing student performance.

For what educational purposes are teachers most commonly using generative AI?

Beyond use of ChatGPT in the classroom, teachers expressed a variety of attitudes regarding the use of generative AI in schools. They emphasized that AI and tools like ChatGPT represent the future of education and foresee a seamless blend of these tools with conventional teaching practices. However, this enthusiasm is tempered by several concerns. A primary issue raised was the credibility of the information provided by the tool. Teachers emphasized the imperative for ChatGPT to source and cite information from reliable and authentic sources, particularly if students are to rely on it for academic research. Another significant concern revolves around the tool's potential for misuse. The ease with which ChatGPT can generate content poses challenges in ensuring students maintain academic integrity, leading to teacher concerns around plagiarism

What do teachers consider to be responsible and ethical use of generative AI in educational settings?

Teachers' attitudes about responsible use of generative AI underscore its potential as a supportive tool, but also emphasize its limitations. While tools like ChatGPT can aid in tasks like content generation and answering questions, educators should view it as a supplement rather than a substitute for their traditional roles. Over-reliance, especially in grading or recommendations, is cautioned against to preserve individualized attention. Personalization of generated content is crucial to align with curricular goals and diverse student populations. Teachers also talked about responsibility of modeling ethical AI use, acknowledging its assistance, and ensuring the authenticity of the information. This balance ensures technology complements, but never overshadows, the essence of teaching.

We are currently running focus groups with students to gather their perceptions and attitudes about generative AI. We hypothetically could see the following trends in student experiences:

1. Students find generative AI useful for immediate clarification: Students may frequently turn to tools like ChatGPT for quick explanations or clarifications on topics they find challenging. However, some feel it could potentially reduce their interactions with teachers or peers.

2. Students emphasize the importance of AI literacy: Students may talk about wanting tp ne educated on the strengths and limitations of AI, ensuring they can use tools like ChatGPT effectively and ethically.

3. Desire for more hands-on training: Students may express interest in hands-on training sessions where they can learn to use ChatGPT and similar tools more effectively, ensuring they harness the technology's potential while adhering to ethical standards.

4. Seeking teacher-student collaboration: Students may suggest that both teachers and students should collaboratively develop guidelines for AI use in classrooms, ensuring a balanced and holistic approach that addresses the concerns and needs of both groups.

Importance

There are several areas where this research can be valuable to ISTE attendees.

1. Informed Implementation: Understanding attitudes allows educational institutions to implement generative AI tools more effectively. Knowing the reservations and enthusiasms of both teachers and students can guide the development of more effective training programs, curricula, and support structures.

2. Ethical Considerations: As AI's role in education grows, it brings along various ethical concerns, from data privacy to the potential for algorithmic bias. By understanding the attitudes of educators and students, institutions can craft guidelines and policies that address these concerns.

3. Predicting Long-Term Outcomes: Teachers' and students' attitudes can provide insights into potential long-term outcomes of AI integration in education. For instance, widespread skepticism among educators might suggest slower adoption rates, while enthusiasm might indicate rapid integration and evolution of teaching methodologies.

4. Promoting Collaboration: By understanding the views of both teachers and students, there's an opportunity to promote dialogue and collaboration. This can lead to co-created guidelines for generative AI's use in schools.

References

Akgun, S., Greenhow, C. Artificial intelligence in education: Addressing ethical challenges in K-12 settings. AI Ethics 2, 431–440 (2022). https://doi.org/10.1007/s43681-021-00096-7

Blose, A. (2023, April 12). As ChatGPT Enters the Classroom, Teachers Weigh Pros and Cons. neaToday. https://www.nea.org/nea-today/all-news-articles/chatgpt-enters-classroom-teachers-weigh-pros-and-cons

Eke, D. O. (2023). ChatGPT and the rise of generative AI: threat to academic integrity?. Journal of Responsible Technology, 13, 100060. https://doi.org/10.1016/j.jrt.2023.100060

Iqbal, N., Ahmed, H., & Azhar , K. A. (2022). Exploring Teachers Attitudes towards Using ChatGPT. Global Journal for Management and Administrative Sciences, 3(4), 97–111. https://doi.org/10.46568/gjmas.v3i4.163

Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., ... & Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and individual differences, 103, 102274. https://doi.org/10.1016/j.lindif.2023.102274

Lo, C. K. (2023). What is the impact of ChatGPT on education? A rapid review of the literature. Education Sciences, 13(4), 410. https://doi.org/10.3390/educsci13040410

Mhlanga, D. (2023) Open AI in Education, the Responsible and Ethical Use of ChatGPT Towards Lifelong Learning. http://dx.doi.org/10.2139/ssrn.4354422

Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A., Hickey, D. T., Huang, R., & Agyemang, B. (2023). What if the devil is my guardian angel: ChatGPT as a case study of using chatbots in education. Smart Learning Environments, 10(1), 15.

Vindrola-Padros, C. (2021). Doing rapid qualitative research. Sage Publications.

Walton Family Foundation. (2023, July 18). Teachers and parents report positive impact of CHATGPT on teaching and learning. https://www.waltonfamilyfoundation.org/learning/teachers-parents-report-positive-impact-of-chatgpt-on-teaching-and-learning