Part 2 - Teaching Assistant

I am choosing to focus on the teacher to parent email.  The teacher asked ChatGPT to write an email to a parent using specific parameters.  In this case, it was about books to a high schooler's parent.  The teacher had to fine-tune her request multiple times to get it where she wanted to email to be.  This included the tone she wanted to email to have and to the grade level she wanted the books focused on.  I have used this tool before to construct an email to my child's teacher.  I did the same type of thing, asking it to be more specific to my needs, until I had what I wanted.  I still added some things and tweaked others, but for the most part I was very impressed with how well it worked. 

The ELA teacher was looking to send an email to a parent about their child who is struggling to complete the reading requirements. She asked ChatGPT to create this email with a few directions; she immediately saw ways that the email could be improved, as it was not specific enough for her liking. Then, she asked ChatGPT to tailor it more for the student's age and interests. She liked this version a lot better, but still found things to remove or change.

In the email to parent example, the ELA teacher was trying to write an email to a parent about a student that is a reluctant reader. She started by giving ChatGPT a request of writing an email to a student’s parent about the book challenge and recommendation. The teacher realized that the original request was too vague. She more specific she got, the better the email was to the parents. It’s important to be detailed in your requests to AI. You can always edit and make it your own. Keep asking it to rewrite it, and give it more specific editing prompts. 

In the email-to-parent demonstration, the teacher typed a clear prompt into the AI tool, explaining the situation she wanted addressed and the tone she wanted the email to have. By being specific in her request, such as asking for a professional but supportive message, she was able to get a draft that matched the response she was looking for. This showed how giving the AI enough context and direction can make its output more useful, instead of vague or generic.

It's also notable that after the AI provided a draft, she reviewed and edited it to make sure the message fit her student’s situation. She adjusted the wording to reflect her own teaching style and added personal touches so the parent would feel the message was genuine and tailored to their child’s needs. This step highlighted how AI can be a helpful starting point, but the teacher’s judgment and personalization make it effective.

     There was a demonstration where the teacher created an email to a parent, she began by clearly stating her goal, communicating a student’s progress in a professional yet supportive tone. She entered a short prompt into the AI tool, outlining what she wanted the email to include, such as the student’s strengths, areas for growth, and ways the parent could support at home. By giving the AI specific details and then adjusting the draft it provided, she was able to quickly generate a polished email that saved her time while still meeting her communication needs.

I focused on the parent email. At first, the teacher wrote more generic details like “write an email about a book challenge.” It came up with things that are a little lower than their grade level and not in the genre students like. After reading the response, the teacher realized they needed to be more specific with details like the grade level and genre of the books. The more details you can provide the more customized the email will be. Another great thing about this is that you can mix and match and pick and choose which parts of the response you would want to use in an email.

I chose to review two of the techniques used in the provided video the parent video and the formative assessment.  

During the portion about the parent email we saw that she started with a basic request "write me an email reply to a student's parent about the book challenge and recommendations". Initial thoughts the teacher had were that the email could be better. Based on the response the teacher realized that additional information (parameters) would be needed to compose the parent email that she wanted so the teacher progressively added more detail and requirements (ie. grade level, genre). Email drafted at this point is of much better quality - the updated version after adding specifics has a better tone however it required final modifications- to modify length, formality and oddly AI removed not just one but two books from the final draft. Final takeaways were that after re-reading the draft made the teacher was able to make final modifications to ensure that the email to parent drafted reflected what the teacher wanted. 

During the portion about the formative assessment the teacher asked for "a five-question quiz based on a website's content to ensure student understanding". She found that the response back from AI didn't include the website which wasn't really surprising. Thankfully the teacher realized that she needed to include a link the website in the prompt. Once it had the website the majority of the questions provided were formatted as short answer. To further drill down the final product the teacher asked that the format of the assessment be moved to 50% of the questions being multiple choice based. AI made revisions and now 50% of the questions were multiple choice - the teacher seemed disappointed as she wanted an answer key and had to ask AI with a separate prompt to provide the answer key.  To further enhance the assignment and confirm student understanding she asked for creative ways to incorporate content and AI created 5 different options of activities that allowed for creativity. 

Overall, we learned that while in our mind the prompts we are giving AI might be too vague. As a result AI leans on it's training because it lacks the common sense that we have as humans. To avoid the disappointment that the teacher had during this video it's suggested that we provide more detail to ensure that we get the final product desired. 

I really liked listening to them using AI to email a parent. I have used AI to help me brainstorm and find the correct wording when messaging a parent. Like they suggested, and had to do, but they need to use specific keywords or clarify some parts. When asking AI to change or add certain things it was quick and precise on what exactly needed changing.  This has been very helpful for me when reaching out to kindergartener parents on day to day things!

In the video one of the demos that stood out to me was when the teacher used AI to make an assessment. She just typed in a clear prompt. Because she gave the AI specific info like the grade level and topi,  it gave her exactly what she needed. It showed me that how you ask the AI, matters. The more detailed you are, the better the AI’s response will be. I thought it would be helpful to do this and have it aligned with the standard it matches as well! 

She first started out with writing a prompt on Chat GPT to write an email reply to a student's parent about the book challenge and recommendations. She then edited the prompt to be more specific about the grade level and the genre that the student would enjoy and have the recommendations to be more at the student's level. Chat GPT was able to customize the response based on her prompt. She read through and decided she wanted to take out a few things such as one of the book recommendations and changed the length of it. She edited the prompt for Chat GPT to do those things and it was able to edit the response to give her what she was looking for. 

The teacher in the video be by telling chat GPT to write an email reply to a parent 10th grade book choices. Fine tuning her perspective allowed the teacher to refine and specify her request. The teacher demonstrated how you can easily alter diction, syntax, length, omissions etc. The more specs you include with your prompt, the more useful and  accurate the result is going to be. I like the idea of using this as a jumping off point for writing letters of recommendation. This could potentially be a huge time saver for periods in the year when I find myself writing a lot fo letters for scholarships, college acceptance, New Visions recs, etc.