Artificial Intelligence in the Classroom: Analyzing the CS Learning Experience in the Context of Generative AI Tools; Investigating the Pushes and Pulls in the Progression of Deepfake Legislation

Over the past two to three years, generative AI has taken the world by storm. From free online gen-AI tools like ChatGPT to AI-based photo editors and even AI-based emojis, we, as a society, have seen an increasing number of sectors become more dependent on artificial intelligence, and generative AI specifically. My technical report was motivated by how generative AI has impacted my life as a college student. As a student who went through at least one semester of college without the widespread use of ChatGPT and then many more with it, I have seen and experienced first-hand a shift in computer science learning as a result of generative AI tools. However, the educational sector was not the only one that experienced a massive change because of gen-AI. Its influence is felt especially in the entertainment industry, where we see AI-generated content on social media, the news, phone apps, and more. Deepfakes, a specific subsection of generative AI piqued my interest, as I saw countless deepfaked videos and pictures circulating the internet. This drove me to think more deeply about how generative AI technology like deepfakes can be misused and how we can prevent that.

My technical report centers on how the widespread use of artificial intelligence tools, such as ChatGPT, has greatly impacted the computer science classroom environment, both for learning and for completing assignments and assessments. To embrace these inevitable changes and acknowledge the proliferation of generative AI tools, I propose implementing curriculum changes that teach students when and how AI tool use is appropriate and helpful in computer science. To define these curriculum changes I propose a method to analyze students’ experiences in both project-based and theory-based courses. Additionally, I propose a method to analyze students’ experiences using AI tools for coding. I expect to find specific use cases of AI tools that can be integrated into aspects of a CS class. I also expect to suggest changes to the current process of administering assessments and assignments that should be completed without the use of generative AI tools. Future work on this subject would encompass seeking feedback from both students and educators on the curriculum ideas developed through this project.

My STS research paper analyzes deepfake technology policy and what makes it successful versus unsuccessful. Deepfakes have many applications in different fields and industries such as education (personalized curriculums), business (upgrading the online shopping experience), and entertainment (recasting actors, changing scene backgrounds, etc.). However, as with any new technology, deepfakes can be used for malicious purposes. Especially with how social media provides attackers with the ability to easily and vastly proliferate their deepfake creations, malicious content like deepfake porn or any non-consensual video of a victim can have profound implications. Often in these cases, victims of malicious deepfakes don’t have clear actions or legal pathways to turn to for remediation. In addition, there is no legislative block at the federal level against this content being created in the first place. Therefore, deepfake legislation is needed and is needed urgently. However, progress in enactment of proposed deepfake legislation has been slow and many acts have been unsuccessful. This can be attributed to a myriad of factors.

Firstly, proposed deepfake legislation lacks focus on the victim and instead often focuses on the deepfake content itself and prohibits it. This makes the legislation too indirect to effectively address victim concerns. In addition, created deepfake legislation can tend to use vague language that is not specific enough to ban specific deepfake use cases or provide a course of legal action. Finally, deepfake legislation often does not define a clear structure for victims to follow when seeking help to hold those liable accountable legally. Successful deepfake legislation that has progressed farthest towards enactment addresses all of these issues and flaws. The findings in this research should aid in working towards better processes and considerations regarding writing technology-focused laws.

Through the process of writing and researching both my technical and STS papers, I was able to reflect more wholly on how generative AI is impacting society and us as individuals. My technical report focused on studying the potential positive impacts of generative AI, implementing these into a classroom, and teaching students when and how to appropriately use it. On the other hand, my STS report focused on tackling situations in which generative AI is inappropriately used. These two topics go hand in hand and result in a more informative and full understanding of how we should approach generative AI tools with openness but also with precaution and a social conscience. Through my research over this past year, I have gained a more profound awareness of generative AI’s role in society - a perspective that I hope to pass on to others.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Rosanne Vrugtman

STS Advisor: MC Forelle