With the rising prominence of Artificial Intelligence (AI) in everyday lives, many districts and schools have raised questions and enacted policies regarding appropriate usage of AI. This study examines student perspectives regarding the usage of AI in school settings.
Problems and Purpose
With the rising prominence of Artificial Intelligence (AI) in everyday lives, many districts and schools have raised questions and enacted policies regarding the appropriate usage of AI. In schools, common uses of AI include using AI as a resource for school work, using AI to help with coding and writing, or relying on AI completely as a tool to complete homework.
Given the sudden rise of AI, especially following OpenAI’s ChatGPT public release in November 2022, some districts and schools have rushed to implement policies against AI out of fear that AI could negatively impact students’ education. However, students were often absent from the conversation, so the following study looks to understand the perspectives of students regarding the usage of AI in United States high schools.
Background History and Context
This study draws on Deliberative Polling.
Organizing, Supporting, and Funding Entities
The Deliberative Democracy Lab at Stanford University: This online event was led by high school research assistants at the Deliberative Democracy Lab. The Deliberative Democracy Lab is devoted to research about democracy and public opinion through Deliberative Polling.
Generation Lab: This event partnered with Generation Lab, which is a leading polling organization for youth aged 13 to 17, in an attempt to gain a more diverse and representative sample of US high school students.
Participant Recruitment and Selection
Participants were recruited through two main methods. First, around half of the participants were recruited through the Generation Lab’s national panel. These participants were more diverse and reflective of the national sample. However, given the financial limitations of the project, the other half of the participants were recruited through the efforts of the Deliberative Democracy Lab’s high school researchers, their schools, and their networks. As such, this deliberation follows the Deliberative Polling methodology, except that the project did not have a representative sample of high school students in the country.
A total of 115 high school students from across the country self-selected to be a part of the deliberation. Participants were incentivized to participate through numerous ways, such as service hours required by schools, a certificate of participation, and the opportunity for a college advice panel at the end of the deliberative exercise. Out of the 115 participants, 71 completed both the pre-deliberation and post-deliberation surveys, so the results are only indicative of the opinions of the 71 who filled out both surveys.
While the participants of the study may not be representative of high school students nationally, the procedures of this event follow standard Deliberative Polling methodology, so results do represent the thoughtful opinions of the participants.
To see more on the demographics of self-selected participants, please view pages 8 to 11 on the Final Report of the study: https://drive.google.com/file/d/1LqfyjbZlm460GfE_GCKBTmZzoWv10i9y/view.
Methods and Tools Used
This online event was hosted on the AI-assisted Stanford Online Deliberation Platform. The Platform is designed for small group discussions and allows for structured and equitable conversations between participants. Read more about the platform at https://deliberation.stanford.edu/tools-and-resources/online-deliberation-platform.
What Went On: Process, Interaction, and Participation
Prior to the Deliberative Polling exercise, participants were sent the Briefing Materials of the deliberation, which introduced participants to AI tools and the proposals that the deliberation would focus on. The Briefing Materials of the deliberation can be found here: https://drive.google.com/file/d/1E9gOnK3sEoHQSj0wo2CmEM-oW3CFwWKH/view.
Throughout the deliberation, 14 proposals were discussed. Proposals 1-7 focused on AI usage in US high schools, proposals 8-10 focused on penalties of unauthorized use of AI in US high schools, while proposals 11-14 focused on AI usage by teachers in US high schools.
Participants responded to the deliberation proposals with a numerical value on a 0 to 10 scale. Those who selected 0 to 3 indicated their strong opposition to the proposal, those who selected 4 indicated their moderate opposition to the proposal, those who selected 5 neither supported nor rejected the proposal, those who selected 6 showed moderate support for the proposal, and those who selected 7 to 10 indicated their strong support for the proposal.
Participants were surveyed twice: once before the deliberation and once after the deliberation. Hence, this study also analyzes the changes in results before and after the deliberations.
Influence, Outcomes, and Effects
Overall, high school students of the deliberation viewed artificial intelligence in a positive light. However, high school students also believe and understand that AI can have a negative impact if not regulated.
Out of the policy proposals proposed, the most strongly supported proposal was “Schools should provide guidelines and resources to teach students how to use AI responsibly”. On the other hand, the most strongly supported penalty proposal was “Students who violate their school’s AI policy should be subject to a warning and/or a grade deduction”.
Additionally, it’s important to note that participants were generally opposed to proposals that resulted in suspension or expulsion for students who violated their school’s AI policy. When selecting appropriate measures of punishment, students preferred corrective penalties rather than punitive penalties.
When it comes to the impact of AI in grading and homework, students showed concerns about detection accuracy of AI detection tools. Particularly, students were concerned with the ability of AI to also exacerbate misinformation– especially if students and educators overly depended on AI to complete assignments.
References
Briefing materials: https://drive.google.com/file/d/1E9gOnK3sEoHQSj0wo2CmEM-oW3CFwWKH/view
Final Report: https://drive.google.com/file/d/1LqfyjbZlm460GfE_GCKBTmZzoWv10i9y/view
