Use of Artificial Intelligence in Education

The Use of AI in Education

Everyone is talking about OpenAI’s ChatGPT and debating about its potential impact on—well everything. And many companies are pouring resources into the research and development of their own language models. While others are already integrating and collaborating with ChatGPT to enhance the user experience and provided updated data and other functions. Open Table, Expedia, Slack, and Shopify are among the first ones.

We’re educators here at Nucamp, but we’re also technologists. And if there’s a cutting-edge product that could benefit us and our students, we can’t help but be fascinated and want to test-drive it and be early adopters.

So, we developed a Discord bot, powered by OpenAI's ChatGPT.

Our motivation for developing the bot was to see if it could help our students learn to code more efficiently, by helping them debug their code.

We affectionately named our AI Tutor, Debugginator.

Starting March 10th, we released the use of the Discord bot available to all of our coding students. The results were interesting.

 

The Birth of Debugginator and Initial Challenges

As with all new development projects that require integration and coordination, we encountered our share of challenges.

1) ChatGPT 3.5 provided the initial code for our Discord bot.

2) But since the data ChatGPT 3.5 has access to is not the most updated, we had to edit the code to use the latest versions of OpenAI’s and Discord’s APIs.

3) We also developed a workaround for Discord’s 2000-character message limit.

4) Our product needed to accommodate attaching files to messages, so we added that enhancement.

After addressing these development issues, Debugginator became a dedicated channel in our Discord server, where students could use the AI Tutor.

The Initial Rollout of our AI Tutor for Our Online Courses

Our students initially struggled with the user interface of the bot. It wasn’t intuitive or user-friendly. People are used to conversational Chat bots where they just type in a question. They may not get the desired answer, but the user interface is simple and intuitive.

Initiating the use of Debugginator required students to mention it before it could be used. They had to provide their code using backticks or as an attachment. And if they forgot any of these steps, they didn’t have a good user experience.

So, we updated the UI to be more intuitive and engaging.

Finding the Right Balance Between Teaching and Learning

When learning anything, but especially when learning coding languages, it’s important to work through a challenge and figure out the problem. Students learn so much when triaging a problem.

We had to find the right balance for the AI Tutor to guide students toward finding the answer rather than just hand them the answer. It needed to provide help, without hindering a learning opportunity.

We continuously fine-tuned the context provided to ChatGPT so it returned answers that weren’t too vague or too specific.

Collecting Feedback about the use of AI Tools in the Learning Process

It was important to measure the effectiveness of the AI Tutor. We did this by surveying the students and asking them to rate the helpfulness of the answers they received on a scale from 1 to 5.

Although we didn’t get 100% of the students to provide ratings via the surveys, some did share valuable feedback about their experience:

• "The debugginator has been incredibly helpful in figuring out exactly what issues exist within my code. Even when the debugginator does not tell me exactly how to fix my code, it gives me enough of a push in the right direction to figure it out myself and thus learn from my mistakes." —Will Augustyn, Nucamp Student
• "I find it quite useful, especially because it seems it has been trained specifically on NuCamp's course content. I usually find that it is helpful if I describe the problem. Sometimes I use it to debug my code for my own projects." —Ryan Mendoza, Nucamp Student

Importance of Human Interaction and Instructor Integration in Teaching and Learning

We’ve always believed that having dedicated instructors and a supportive community are invaluable for the learning process.

We were curious if having access to a tool such as AI Tutor would impact the amount of collaboration between coding students and their instructors.

But we observed that the amount of person-to-person interaction didn’t diminish. Students continued to engage with their peers and instructors.

The AI Tutor was another tool they could use to help with efficiency and research. It saved them time searching for answers online.

This reinforced our belief that human connection in the learning process is irreplaceable.

Also, our instructors tested the capabilities and limitations of the AI Tutor.

They saw it as an effective tool to supply targeted guidance different from what students would find doing an online search. It provided debugging guidance but was no replacement for human instruction.

Measuring the Impact of using AI Tools in Student Learning

It’s too early to measure whether the use of AI in education will change student progress or success rates. We need more data. But we plan to continue tracking these metrics to better understand its impact on the learning experience.

Academic Integrity: Addressing Ethical Concerns

Researching problems is integral to the learning process. As stated above, we took measures to ensure the AI Tutor did not give answers to the students, but it provided guidance to help them discover the answers independently. This allowed students to learn without compromising academic integrity.

The Future of AI in Education and the Competitive Advantage of using AI tools

We’re excited about the growing potential Debugginator has for enhancing the student experience. We’re working on a version allowing students to interact with the bot verbally, providing a more natural and engaging user experience. We plan future enhancements to include personalized learning for struggling students, assisting with personal projects for programming portfolios, helping them prepare for job interviews, and expanding the AI Tutor's capabilities to cover more curriculum areas in our coding courses.

Currently, we view the Debugginator as a competitive advantage for our coding bootcamp. It sets us apart from other learning institutions by offering cutting-edge AI tools. However, we do not intend to seek partnerships at this time. We believe that the AI Tutor's unique benefits can help us continue to attract and retain students who are eager to learn to code in a supportive and technologically advanced environment.

Key Takeaways:

Our experience with the AI Tutor has been a valuable lesson in the potential and limitations of using Artificial Intelligence in education. While the Debugginator has saved our coding students time and provided targeted guidance, it has also reinforced the importance of human interaction in the learning process. As we continue to refine and expand our AI Tutor, we remain committed to fostering a balanced learning experience that combines the best of both worlds: the efficiency of AI-powered tools and the irreplaceable value of human connection, support, and collaboration.