🏫 Illusion of Mastery with AI
Explore the Impact of AI on Skill Mastery in Education
A recent issue of StudentPOLL featured the results of their survey of over 1,300 high school seniors in the United States, revealing how students have been utilizing AI for academic work and college searches. Some interesting findings from the research include that over a third are using AI for schoolwork, while only a small number are leveraging it for their college search processes. According to the survey, students also expressed concerns about the ethical implications of generative AI and expect colleges and universities to educate future students on the effective and ethical use of these tools. Below are some of the charts featured in the Volume 17, Issue 1 of studentPOLL by Art & Science Group, titled, “AI and Academia: Student Perspectives and Ethical Imlications”:
If you're interested in exploring the latest discussions about AI within educational settings, join us at AI x Education for a new series of free webinars. These sessions will feature insights from educators, researchers, and students from diverse backgrounds.
In our upcoming webinar on May 27th, we will be featuring Eric Klopfer, a renowned professor at MIT specializing in educational technology and game-based learning. Register now through this link!
Here is an overview of today’s newsletter:
Popular AI tools for academic research
How an aerospace engineering student leverages AI for coursework and time management
Research study on educators' ability to detect AI-generated content
The World Economic Forum's Report on "The Role of AI in Education 4.0"
🚀 Practical AI Usage and Policies
It can be tedious to sift through hundreds of research papers to find relevant and high quality resources for academic research. However, with the help of AI-powered tools, this process can be streamlined. These tools can serve as an advanced search engine and offer additional capabilities such as citation mapping, summarization, data visualization, and more. Here are some popular AI tools being utilized by students and researchers for academic research.
Semantic Scholar
Semantic Scholar is a free resource developed by the Allen Institute for Artificial Intelligence that uses AI to help users navigate the vast amount of scientific literature available. With over 200 million papers across all fields of science, Semantic Scholar offers a comprehensive database for research exploration. They also offer an augmented reader called Semantic Reader that highlights key points, offers in-context explanations, and reveals connections to other relevant research within Semantic Scholar's vast database.
Scholarcy
Scholarcy is an AI tool designed to summarize, analyze, and organize research. It summarizes research papers, articles, textbooks, and videos, and converts them into interactive summary flashcards highlighting key information. Users can keep their research papers organized into folders and synthesize their insights.
Scite.Ai
Scite.Ai is an AI-powered research tool that focuses on evaluating the credibility of scientific claims. It analyzes the text of research papers, using deep learning algorithms and natural language processing, to understand how they cite each other. They offer “Smart Citations," which reveal the context behind citations. This allows users to see if a cited paper supports, contradicts or simply mentions the claim in question. By offering this deeper insight into the relationships between research papers, Scite.Ai allows users to assess the validity of scientific arguments and navigate the intricate web of scientific knowledge.
Elicit
Elicit is an AI-powered tool that allows users to pose research questions and instantly retrieve relevant papers from its vast database. It also provides one-sentence summaries to help users prioritize, and meticulously organize extracted details into clear tables. This allows users to identify themes and concepts across multiple papers with ease.
Consensus
Consensus is an AI-powered research assistant designed to tackle scientific literature. It utilizes artificial intelligence to understand the content of research papers and pinpoint the ones most relevant to your inquiry without having to keyword match. Unlike a traditional search engine, Consensus prioritizes credible sources by focusing on peer-reviewed research. Users can search by keywords, topics, or through direct questions.
Connected Papers
Connected Papers is an AI-powered research navigator that can act as a visual springboard for research exploration. It takes a single research paper as the starting point and leverages AI to generate a visual map of connected research. This map displays other papers related to your initial one, with the size and color of each node indicating its significance and publication date. By exploring these connections, users can gain a broader understanding of the research landscape and discover new, relevant papers they might have missed.
📣 Student Voices and Use Cases
This week, we spoke with Jackson Frazier, a rising third-year student at Embry–Riddle Aeronautical University majoring in Aerospace Engineering. In the following, we present select highlights from these conversations, which have been slightly edited for enhanced clarity and precision:
Q: In what ways have you used AI in your studies?
I used AI tools to help me study for some of my finals. I had already done a lot of the problems we were given in class, so I already knew how to solve them. I wanted some stuff I hadn’t really seen before, so I asked ChatGPT to generate me problems based off the reference examples from my textbooks. I would upload screenshots of example questions and it would generate similar questions I had never seen before which helped me to study. I find that when I’m studying, if I do the same problem over and over again, I kind of remember how to do that one specifically, as opposed to the process. With this method, it helped me fortify the process of how to solve these unique problems that I've never seen before. I also found that some of the questions were better than others, but the complex ones were especially nice because those helped me to understand the concepts a lot. I would also use ChatGPT to help me when I’m stuck on solving the question, especially in explaining those gray areas where I don’t feel confident in my understanding of it.
Q: What are the biggest benefits and challenges you face when using AI for learning?
One of the biggest issues I face with using AI is that it tends to get some stuff wrong, and obviously, I'm not perfect, so I can't always realize that. I've had times where it taught me how to do the process the wrong way. That’s something that comes with new technology and the risk you take using it. When that stuff does happen, I'll go back afterward and in the chat and say that it was wrong in order to correct it.
One of the biggest benefits I found was that AI tools are very helpful for my understanding of new material. Some teachers will explain stuff in a way that assumes you already have a master's degree in the topic, but with ChatGPT, you can tell it to simplify complex material. If you still don't understand, you can ask it again until you get a version of the explanation that you understand. It can also simplify the professor’s words. Sometimes, I will take transcripts from my class lectures and then have them simplified for me. I found this to be a game changer, especially for my harder classes.
Q: How have you used Generative AI tools outside of your coursework, if at all?
I use ChatGPT to generate Google calendars for me. One of my friends showed me how I can connect it to my Google calendar to give me a layout of the day. I found this to be a lot better than just using reminders on my phone because I feel like whenever I have that, I don't always listen to it. Now, I can have these notifications pop up on Windows, and it will remind me, “Oh, let me go and get this done really quick, or I need to go do this right now.”
The way this system works is that first I will tell ChatGPT about myself, like "I'm Jackson. I'm an aerospace engineering major. These are my hobbies, and this is what I want to achieve." Then, I'll give it a list of things I want to achieve. I'll also import my class schedule and other goals. After that, it will give me a layout of a basic day plan of what to do. I've been using this system a lot, and it has helped me become much more organized as a person, which was something I struggled with.
📝 Latest Research in AI + Education
Computers and Education: Artificial Intelligence
Do teachers spot AI? Evaluating the detectability of AI-generated texts among student essays ↗️
This paper examines whether novice and experienced teachers can distinguish AI-generated text from student-written essays. The study reveals that both novice and experienced educators struggle to accurately identify texts produced by AI, specifically those generated by ChatGPT. Teachers often misjudge their ability to discern the origin of the texts, displaying notable overconfidence in their assessments. The experiments, involving 89 novice teachers and 200 experienced teachers, show that AI can adeptly simulate student writing styles, making detection challenging for teachers. Additionally, it was found that AI-generated essays are sometimes rated more favorably than those written by students, which could have significant implications for educational practices and assessments.
Fleckenstein, J., Meyer, J., Jansen, T., Keller, S. D., Köller, O., & Möller, J. (2024). Do teachers spot AI? Evaluating the detectability of AI-generated texts among student essays. Computers and Education: Artificial Intelligence, 6, 100209. https://doi.org/10.1016/j.caeai.2024.100209World Economic Forum
Shaping the Future of Learning: The Role of AI in Education 4.0 ↗️
This report by the World Economic Forum delves into the integration of artificial intelligence (AI) in education, emphasizing its potential to revolutionize teaching and learning under the framework of Education 4.0. This framework, developed by a coalition of global experts, aims to equip learners with the skills and competencies needed for the future. The report highlights how AI can significantly enhance personalized learning, automate administrative tasks, and improve assessment accuracy, thus allowing teachers to concentrate on more personalized and impactful teaching methods. This will help bridge the gap between the increasing demand for teachers (an additional 44 million will be demanded by 2030) and the already exhaustive workloads faced by educators. However, the paper also addresses the challenges and risks, that currently lead 71% of people to be concerned, such as the potential for AI to disrupt traditional educational roles and the need for proper safeguards to manage data privacy and ensure equitable access to technology.
Daepp, M. I. G., & Counts, S. (2024). The emerging AI divide in the United States. arXiv preprint arXiv:2404.11988.📰 In the News
IEEE Spectrum
AI Copilots Are Changing How Coding Is Taught ↗️ (Featuring AI x Education’s team)
Key takeaways:
Generative AI is increasingly used in academic settings, particularly in computer science education, where it assists students in understanding complex concepts and enhances their coding skills. This integration is prompting educators to revise their teaching methods, focusing less on syntax and more on higher-level cognitive skills like problem-solving and testing.
Educators like Jeanna Matthews from Clarkson University are exploring ways to incorporate AI tools in the classroom while ensuring that students still learn foundational computer science principles. These tools are seen as a way to enrich the learning experience rather than merely simplifying the coding process.
Teaching strategies are shifting towards more practical applications, such as having students test AI-generated code for correctness and engage in group projects that emphasize the software development life cycle. This approach aims to prepare students for real-world software engineering roles by encouraging collaboration, design, and ethical considerations.
While the benefits of using AI in education are recognized, there are concerns about students becoming overly reliant on AI, potentially undermining their learning process. Educators stress the importance of using AI as a "copilot" to supplement learning, not as an "autopilot" that diminishes critical thinking and problem-solving skills.
Inside Higher Ed
Academic Success Tip: Infusing AI into Curricular Offerings ↗️
Key takeaways:
Generative artificial intelligence (AI) is being integrated into higher education curriculums to prepare students for the workforce, with 14% of chief academic officers already reviewing curriculums and 73% planning to do so. This reflects a significant trend towards future-proofing students' skills for a technologically advanced job market.
Faculty members are leveraging AI in various educational roles: as a tool to enhance learning outcomes, as a topic of direct study within courses, and as a component in research. AI tools like chatbots are increasingly used to provide both academic support and emotional encouragement to students.
Several universities, including the University of South Florida and the University of Georgia, are implementing AI-driven chatbots in courses to help with academic queries and motivational support. This is part of broader initiatives to explore the impact of AI on student outcomes.
Courses specifically on AI are also being developed; for instance, the University of Mary Washington and American University are teaching students how to build and utilize AI tools, with American University requiring new students to complete AI-related courses as part of their core curriculum starting in fall 2024.
Research-oriented AI integration is also being emphasized, with institutions like Cornell University and Boston University exploring AI's role in research and academic writing, aiming to establish best practices and ethical guidelines for its use.
“Chatgpt.” ChatGPT, OpenAI (GPT-4), openai.com/chatgpt. Accessed 13 May. 2024. And that’s a wrap for this week’s newsletter! Based on our previous poll, we found that a majority of our readers use GPT-4 most frequently. According to Open AI’s latest announcement, GPT-4o (“o” for “omni”), a better model than GPT-4, is now available for free for all users!
If you enjoyed our newsletter and found it helpful, please consider sharing this free resource with your colleagues, educators, administrators, and more.