Artificial Intelligence in Education

Background & Objectives

Artificial Intelligence (AI) is becoming increasingly common in various aspects of life, especially in education. Research has shown that student awareness in AI is growing quickly both domestically and internationally. Many companies are quickly launching various educational tools that include AI capabilities. This research aims to understand student perceptions and behaviors around ChatGPT. Additionally, this research will gather feedback on various AI concepts for students:

• Current ChatGPT usage

• Primary benefits and challenges of using Chat

• GPT for undergradAccuracy of trust of ChatGPT

Method

Stakeholders wanted to hear from students on both sides of the spectrum that were early adopters, ranging from frequently used to light users.

I worked with our PM to define the different user types that our stakeholders want to learn from. I socialized these definitions to our stakeholders for participant buy-in. We wanted the range so that we can see the different use cases that users might be using chatGPT. Why do some students become power users, while others may stay light users?

I conducted moderated interviews with undergraduate US college students that were either power users or light users.

Participants

Definition

• Power users – Students that use ChatGPT for educational purposes 3-5 times a week

• Light users – Students that use chatGPT 0-1 times per week

5 undergraduate students located in the US that indicated usage of ChatGPT for STEM-B and writing classes

• 3 power users, 2 light users

Analysis

This was a qualitative study with many different observations, so I chose to use affinity diagrams to analyze the user interviews. This allowed me to put together research-based use cases for how students use artificial intelligence in an educational setting. Each note came from a direct observation from transcripts, which allowed me to interpret and identify the significance of each note. From the affinity diagram, a story emerged about the different use cases and the nature of their problems.

Outcome

Once this project became top of mind for our stakeholders, all the way up to our executives, I started with a preliminary report to share with stakeholders. The preliminary report was to provide early information to our stakeholders to have educated conversations. A robust report of use cases was put together and shared out for reference a week later. This was an in-depth report that informed our stakeholders of what they should expect from students and their AI usage.

Impact

This study steered the way our executive team made decisions about the future of AI and Chegg. It included the different use cases, benefits, challenges and accuracy of tool that drove conversation for decision making. There were clear patterns of how students use ChatGPT in education.

Below are a few of the main ways that students use ChatGPT:

• Student expectations begin low, but adjust as they gain experience

• Students use ChatGPT to understand and explain difficult concepts in a simple way

• ChatGPT helps put students on the right path or verify they are answering correctly

• Students use ChatGPT as a starting point when writing papers

Reflection

There is a sweet spot for compromise with leadership: It’s common for teams to be excited about research findings, and would set immediate deadlines. It’s important to meet deadlines, but it’s also important to set expectations to ensure that we have adequate time to process research data. While working with cross-functional partners, there is always an opportunity to pushback on timelines if you present the appropriate reasoning.