Problem
As a first-generation student who served as a peer mentor for six years, I wanted to develop an AI platform that could interact with students and provide strength-based mentorship. The goal was to test the capacity of strength based feedback from platforms such as ChatGPT or Forefront.ai and determine if they could provide culturally and grade-appropriate feedback for marginalized learners. The challenge was to create an AI prompt that could understand the nuances of human behavior and provide personalized and peer driven guidance.
Learn more about why strength-based mentorship matters for first-generation students here.
Process
Strength-based mentorship is defined as growth feedback that focuses on unique and specific assets of students and leveraging those skills to make up for lack of others in varying situations.
I chose Forefront.ai as my platform as I aimed to develop a prompt that could interact with students in a natural language format. I used a guided prompt rational that descriptively laid out the AI's responsibilities, along with setting the framework for when it would pause to access situational responses and student interactions, which allowed it to learn from past experiences and improve its responses over time.
Results
AI had a success response rate of 100%, meaning the prompt did not crash the platform or ask it for clarification
AI was able to interact with the user for 10 minutes until responses were being repeated
AI gave general academic advice and peer guidance based on situational awareness and was able to continue conversation while taking necessary pauses
Implications
Lacked direction and specificity
The AI was regurgitating generic advice twice before I asked it to stop. I think the general issue with using AI in education is that it can truly never understand human contexts on a sophisticated level, and thus wont be able to provide feedback. I can see AI being beneficial to people who are using it for metrics and data, but for more lived experiences and feedback may be a barrier because they are not humans who possess the cultural and personal awareness on how to provide mentorship for students.
Questions to consider:
1. What are the capacities of AI feedback and specialized mentorship?
2. Can intelligent computer systems truly access cultural and situational awareness to the full extent a human mentor could?