The new world: artificial intelligence in education
21 Nov 2023
Introduction
In recent years, artificial intelligence has become a large part of education, learning, and software development. From predictive text to large language models, almost every student has utilized or been exposed to some form of AI. I have not used it extensively myself, but the main reasons I see it being used are for information gathering, summarizing, and for revising writing. In software engineering, this has relevant applications in identifying syntax errors, debugging, and identifying suitable functions or features for a task. It is interesting to think about how the use of AI for these purposes influences the type of learning we experience.
Personal Experience with AI in ICS 314
- Experience WODs e.g. E18 - I have not used any form of AI for Experience WODs. I did not encounter any problems that could not be solved using simple Google search or using the resources we were given.
- In-class Practice WODs - I have not used any form of AI for in-class practice WODs. Though it probably would have helped me identify errors and debug, I was still unfamiliar with ChatGPT and Github CoPilot, so I did not think that I would be able to utilize them efficiently.
- In-class WODs - I have not used any form of AI for in-class WODs. Similar to the in-class practice WODs, I did not feel like they would be useful to me. I was also able to solve most of the WODs without AI, so I became more confident in my own unassisted ability.
- Essays - I have not used any form of AI for essays. I feel confident in my ability to write and check grammar and spelling, so I do not feel the need to.
- Final project - I have not had an opportunity yet to use AI for our final project. Now that I have used ChatGPT, I may be more open to utilizing it for our final project if I feel the need to learn more about a specific component or determine tradeoffs between components.
- Learning a concept / tutorial - I have used AI to help me understand concepts for other classes, but not in ICS 314. I feel satisfied with the information I get through watching tutorial videos, looking through documentation, and reading examples.
- Answering a question in class or in Discord - I have not personally used AI to answer a question in class or on Discord. However, my tablemates used ChatGPT to pick a group name for us at the beginning of the semester.
- Asking or answering a smart-question - I have not used AI to answer a smart question for ICS 314. If I have a question, my first instinct is to search for documentation or search for my issue on Stack Overflow.
- Coding example e.g. “give an example of using Underscore .pluck” - I have not used AI to generate coding examples. I feel that the examples provided in documentation are sufficient, and if I want to see more examples I will use Google.
- Explaining code - I have not used AI to explain code. I do not think I’ve had an instance in this class where I was unable to understand code. If I struggled, I would likely just refer to the documentation.
- Writing code - I have not used AI to write code. I know that ChatGPT and Github CoPilot can generate code blocks, but I feel like this would slow me down if they used a component I was not familiar with or if I had to debug that code.
- Documenting code - I have not used AI to document code. I feel that I can do this sufficiently on my own by writing my own comments and keeping a neat style.
- Quality assurance - I have used AI for quality assurance in other classes, but not ICS 314. I have not felt so frustrated that I felt the need to do so, but now that I have used ChatGPT to analyze code for other classes, I would feel comfortable using it for ICS 314 if I had to.
- Other uses in ICS 314 not listed - I have not used AI for any other purposes in ICS 314. I was not familiar with popular AI tools until relatively recently, and I never felt the need to utilize them even though I knew other students were doing so.
Impact on Learning and Understanding
Overall, the greatest impact that AI has had on my learning and understanding of new concepts is that it has made the process faster. If I ask ChatGPT to recommend functions or algorithms for a specific task, that gives me a branching point to do more research on those specific structures. While I do research, I am more motivated knowing that I am likely on the right track with what I am learning about. I feel more efficient and confident if my guesses for concepts are backed up by the suggestions that ChatGPT makes.
The way that each person chooses to use AI makes a big difference in how their learning is affected. Personally, I have experienced using AI to debug or give me ideas as starting points for further research that I do on my own (without asking more from AI). As a result, when I give prompts to ChatGPT, I need to begin with a solid understanding of my assignment and what kind of results I want to get. In a way, it is similar to using a search engine. I feel that I have grown better at asking specific, relevant questions.
Practical Applications
I think that AI as a tool greatly expands the possibilities of software engineering, both in a sense of capability and in a sense of the developmental process. I have seen AI used in other projects for predictive modeling, navigation, information gathering, image detection and processing, and chat bots. I think that the use of AI makes these features much more useful and easily integrable than if they were implemented by hand. Using AI cuts down on human error and tedious mathematical calculations. However, in all cases it must be responsibly trained, which presents a unique challenge to software engineering.
Challenges and Opportunities
Since I have not used AI in ICS 314 (yet), I am not sure of challenges I would encounter. In other classes, using ChatGPT has sometimes proven difficult because it will make mistakes and sometimes contradict its own information. Even though ChatGPT is efficient for retrieving information, I trust it less than when I use the Internet to search for information myself. I can quickly identify if a feature or an example is relevant to me in the scope of my code, but ChatGPT only has access to what I feed it, which is usually a code snippet of the problem area.
I think that AI has become a permanent part of our daily life, which affects the way students learn, and it would be unwise to completely restrict the use of AI in a classroom setting. I have found that it is a popular opinion among older generations that using AI hinders critical thinking. I think that this can be true depending on how it is used by the student. Copying an entire problem to ChatGPT can yield wrong or inapplicable answers. On the other hand, asking targeted and direct questions can spur further interest and guide the student to applying the information to their unique situation. Learning how to properly utilize AI hand-in-hand with traditional learning techniques is something I believe should be integrated into the classroom.
Comparative Analysis
Traditional teaching methods and AI-assisted approaches to learning inarguably train different skills. I believe both will be utilized in a classroom setting, with a focus on traditional teaching methods for learning new concepts and a focus on AI-assisted learning for checking understanding and writing and debugging code. Admittedly I have been out of touch with the student perspective on how AI should be integrated into the classroom, but I am aware that many students do use AI for learning (even outside of class) and find great value in it. I have personally seen a decrease in classroom engagement because of this, but I think it also depends on the student. A student who prefers independent learning will utilize AI tools more than a classroom lecture, which results in less classroom engagement but possibly a better learning experience for them. In developing technical skills, I think that students who learn to utilize AI for their assignments and projects will have an end-result advantage over students who do not use AI. This is because the workforce values productivity, and as AI advances it will continue to become more and more profitable for a business to utilize AI. A student who knows how to use it will be more successful.
Future Considerations
Some things I am looking forward to seeing in the future other than the obvious advancement of AI to take on more complex tasks are ways in which it will be regulated. Usage in the classroom should probably be regulated to limit “cheating” (in the traditional classroom sense). Also, I think schools should teach ethics related to the use of AI. Given that models need to be trained on data, which can be considered private intellectual property, it is extremely important for AI users to be aware of what is considered ethical use of information and for safety standards to be set. I think that this topic is not well-explored at the moment, but I am looking forward to seeing how this affects AI in education.
Conclusion
In conclusion, AI has taken a firm hold onto the way we teach, learn, and operate about our daily lives. As it continues to grow and we continue to learn how to utilize it well (both in the practical and ethical senses), it is important to be aware of the changes it will make to schools and the workforce. The opportunity to utilize AI in this course is valuable, and though I did not take advantage of it, I recognize the importance of learning to use AI and the unique skills that using AI train.