Event Information
Blended ContentOther presentations in this group:
GenAI tools can be both transformative and challenging for educators and students alike. Experts at MIT point out that “AI can be shockingly good and then comically bad at very similar tasks” (Klopfer et al., 2024, p. 3). Recent advancements in large language models have enhanced GenAI's ability to process human language and generate text, images, audio, and video. Since the release of ChatGPT in November 2022, K-12 students have been using it regardless of whether GenAI has been encouraged or banned at school, while teachers have struggled to adapt. The use of GenAI in education has raised concerns about authorship, learner agency, misinformation, and bias, among others (Han et al., 2024). To educate middle schoolers on the opportunities and challenges of using GenAI, researchers developed online instruction that helped students learn technical information about GenAI models, as well as applications, benefits, harms, and ethical considerations (Ali et al., 2021).
Research about various perceptions and uses of GenAI in education has begun to emerge. A survey by Hopelab et al. (2024) found that K-12 students have a positive outlook on AI's potential to transform information access and generate ideas for academic use. They are also excited about the creative applications of AI in education and beyond. However, Kim et al. (2023) described several “naive conceptions” that middle school students held about GenAI such as (a) AI is automation and robotics; (b) AI is a panacea; (c) AI is smart; (d) AI can use all data; and (e) AI is unrelated to ethical considerations. In another study, researchers found that high school students used ChatGPT for tasks such as generating information, exploring ideas, and completing routine tasks; they also utilized strategies such as refining prompts, checking facts, and evaluating output (Chen, et al., 2023). The students recognized the strengths and limitations of GenAI, such as its ability to generate content as well as its lack of accuracy and transparency. These preliminary findings highlight the need for continued research and education about users’ perceptions and uses of GenAI in educational settings.
We used a mixed methods approach to answer the research questions, combining quantitative data (from surveys) and qualitative data (from focus groups) to provide a comprehensive understanding of middle schoolers’ perspectives on using GenAI.
Participants and Context
A free week-long summer civics camp was held at a private university located in a large urban area in the Southwest. Two main strategies were used to recruit racially and economically diverse students for the camp: (a) inviting past participants to attend and (b) asking middle-school principals, counselors, and teachers to nominate potential participants. The camp roster included 49 students who would be entering grades 6–9 in the upcoming school year. (Six students did not provide consent or assent for participation in the study.)
The demographic and educational backgrounds of the camp participants varied widely. About 74% of the 43 participants were new to the camp, while the remaining 26% had attended previously. Most campers (74%) attended public schools, while 16% attended private schools and 9% were homeschooled. They were entering grades 6-9 as follows: 28% were 6th graders, 30% were 7th graders, 26% were 8th graders, and 16% were 9th graders. Campers identified their gender as follows: 40% female and 60% male. They also identified their race/ethnicity as White/Caucasian (35%), African American/Black (33%), Hispanic or LatinX (14%), Bi-racial (12%), Asian (2%), Native American (2%), and unknown (2%). About 19% of parents indicated their children were eligible to receive free or reduced lunches at school, based on income, the rest were either not eligible, were unsure of their eligibility, or did not answer the question.
Data Collection and Analysis
On the second day of camp, students responded to a 10-item survey. Later that day, the first author presented a media literacy lesson that provided an overview of the research process (planning, searching, evaluating, and citing sources), including tips on traditional internet searches as well as using GenAI for research.
On the last two days of camp, students answered another 10-item survey. In addition, the oldest students (rising 8th and 9th graders) participated in focus groups, so that we could gather more in-depth information about their perceptions of using GenAI. One focus group consisted of six students and the other included seven students.
Data analysis strategies include qualitative coding and thematic analysis (to identify patterns and themes in the interview transcripts) as well as calculation of descriptive statistics and creation of charts and graphs (to summarize survey responses).
While we are still in the process of analyzing the results, we provide a summary below of preliminary findings from survey responses followed by our initial impressions from an analysis of focus group transcripts.
Quantitative Results
Survey responses revealed patterns of student engagement with GenAI, including how frequently they used it, their purposes for using it, and their perceptions of its potential benefits and risks. Responses to the first survey revealed that nearly 60% of students used GenAI once in a while, approximately 15% used it once a week or more, and about 20% had never used it. Almost half reported that they used GenAI for entertainment. Some students also used it for schoolwork in or outside of class. More than half indicated that they had figured out how to use GenAI on their own. Others indicated that they learned from friends, teachers, parents, siblings, or internet sources. Almost half indicated that they would like “a little help” from adults in using GenAI tools. The majority of students thought that GenAI could allow them to get the information they needed and help them learn things they needed to know in the future. About a quarter or fewer agreed that GenAI could make it easier to cheat, create issues with teachers, spread misinformation, or hinder their ability to solve problems independently.
Pre- and post-survey responses showed a notable increase in students' knowledge of GenAI, along with slight shifts in their perceptions of its accuracy and trustworthiness. Students' knowledge of GenAI improved, with a 20-point increase in those reporting they knew a great deal about it and a corresponding 20-point decrease in those saying they knew very little or nothing about it. Their perceptions of the accuracy and trustworthiness of information received from GenAI shifted slightly, with most students continuing to believe that GenAI was somewhat accurate and trustworthy. However, fewer students viewed AI-generated information as inaccurate, and more considered it trustworthy.
Post-survey responses revealed that while most students saw GenAI as helpful for research projects, many also expressed concerns about its potential drawbacks and were uncertain about its future use for school assignments. Specifically, nearly all the students thought GenAI was either very or somewhat helpful for research projects. In addition, the majority thought they would use it for generating ideas, creating an outline, finding information about their topic, and getting suggestions to improve a project. However, most students were concerned that using GenAI for research could pose problems such as getting incorrect, biased, or incomplete information, having difficulty verifying the information, using it to cheat, and becoming lazy or dependent on it. Finally, 35% of the students thought they would use GenAI for school assignments in the future, 37% were not sure if they would use it, and 28% did not plan to use it for school assignments.
Qualitative Results
Data from focus group transcripts indicate that GenAI was perceived with skepticism. None of the 13 focus group participants believed GenAI should be trusted outright, and students in both focus groups identified more drawbacks to GenAI than benefits. Perceived benefits of GenAI included producing creative ideas about topics, providing quick assistance when needing to meet deadlines, and offering information in accessible formats. One camper described using GenAI previously to help him understand information from a book he was reading. When he asked it to “explain it like [he] was a 3-year-old” he noted that “yeah, that worked.”
Drawbacks fell into two major categories: flaws with the technology itself and disadvantages for users their age. Focus group participants identified several limitations of GenAI, including bias, outdated information, misinformation, and inconsistent quality. Disadvantages for users (e.g., participants and other people their age) included “stunting” their learning and creativity and making them “lazy” and over-reliant on technology. Additionally, participants drew attention to ethical issues with GenAI and its use. Concerns were expressed that “AI takes credit for other people’s work” and fails to credit sources. Furthermore, users could engage with GenAI to cheat, robbing them of a sense of accomplishment. One student explained, “[Using AI] wouldn’t feel sincere. It wouldn’t feel like it was from you.”
When asked if ChatGPT might help them finalize their action civics projects, responses were mixed. Both groups had already used a search prompt provided during the media literacy lesson to investigate their topics with ChatGPT, and both found the tool informative for “digging deeper” into content. After the media literacy lesson, one group studying animal abuse used ChatGPT to ask a question about, as one student explained, “where we could find better statistics, because that's something that we needed to add to [our project].” Students reported that Chat GPT recommended helpful websites for additional statistics about their topic.
However, healthy skepticism was also evident as participants considered using ChatGPT for final components of their projects, which included an original PSA video and a Google slide presentation with images and information. One participant noted, “I just feel like we’re at a stage [of our project] that’s more creativity-based. And AI isn’t really good at creativity.” Someone else expressed reluctance to use ChatGPT for their final projects because “It could either give us false information or true information.” Another group member echoed, “It's like this is a very serious topic. And all the information should be 100% true.”
This study contributes to a growing body of research that seeks to understand students’ perceptions and uses of GenAI and determine effective methods for incorporating GenAI into K-12 educational settings. It provides insights into middle school students' perceptions and use of GenAI tools, within the context of a camp focused on civic engagement and digital literacy. By exploring how students interacted with and perceived these tools, the study highlights the importance of teaching students about using GenAI tools for research and reveals concerns related to accuracy, trustworthiness, dependency, and ethical issues.
Realizing that many middle school students have explored GenAI on their own or with some help from adults or peers, we recommend that teachers provide guidance on not only how to properly use the tools, but also when it is appropriate to use them, and for which purposes the tools may be most useful. Instruction on using GenAI tools must also address the accuracy and trustworthiness of information obtained through GenAI, as well as ethical matters surrounding the use of these tools. Since widespread use of GenAI tools in educational contexts seems inevitable, ongoing research is needed to identify effective applications, address implementation issues, and evaluate the impact these tools have on educational outcomes
Ali, S., DiPaola, D., Lee, I., Hong, J., & Breazeal, C. (2021). Exploring generative models with middle school students. Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, 1–13. https://doi.org/10.1145/3411764.3445226
Chen, B., Zhu, X., & Díaz del Castillo H., F. (2023). Integrating generative AI in knowledge building. Computers and Education: Artificial Intelligence, 5, 100184. https://doi.org/10.1016/j.caeai.2023.100184
Han, A., Zhou, X., Cai, Z., Han, S., Ko, R., Corrigan, S., & Peppler, K. A. (2024). Teachers, parents, and students' perspectives on integrating generative AI into elementary literacy education. Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). ACM Digital Library. https://doi.org/10.1145/3613904.3642438
Hopelab, Common Sense Media, & The Center for Digital Thriving at Harvard Graduate School of Education. (2024). Teen and young adult perspectives on generative AI: Patterns of Use, Excitements, and Concerns. https://www.commonsensemedia.org/sites/default /files/research/report/teen-and-young-adult-perspectives-on-generative-ai.pdf
Kim, K., Kwon, K., Ottenbreit-Leftwich, A., Bae, H., & Glazewski, K. (2023). Exploring middle school students’ common naive conceptions of artificial intelligence concepts, and the evolution of these ideas. Education and Information Technologies, 28(8), 9827–9854. https://doi.org/10.1007/s10639-023-11600-3
Klopfer, E., Reich, J., Abelson, H., & Breazeal, C. (2024). Generative AI and K-12 education: An MIT perspective. https://doi.org/10.21428/e4baedd9.81164b06
Back
Trips and Tours