Step 1: Introduction to APIs and AI Tools (15–20 min)

1.     Explain the concept of an API:

o   Use a simple analogy: “An API is like a waiter in a restaurant. You (the user) place an order (your prompt), the waiter (API) takes it to the kitchen (AI service), and then brings back the dish (AI response).”

o   Mention that educators can use APIs to make learning tools smarter—like automatically generating definitions, summaries, or quiz questions.

2.     Show examples of API use in education:

o   Translation apps (Google Translate uses APIs)

o   Chatbots for student support

o   Automated writing feedback systems

3.     Explain the tool you’ll use today:

o   Google Colab — a browser-based coding platform (no installation needed).

o   OpenAI API — the “brain” that will generate text responses.

Trainer’s tip: Keep the explanation light and interactive. Reassure participants that no coding experience is required—they will copy and modify existing code.

Step 2: Access the Template (10 min)

1.     Trainer shares the Google Colab link (e.g. a prepared notebook such as) https://colab.research.google.com/drive/1gTexampleLink

2.     Participants open it in their browser (must be logged into a Google account).

3.     The notebook has three prewritten cells:

1.     Install and import libraries

2.     Enter API key

3.     Send a prompt to the AI model

Trainer shows how each cell can be executed using the ▶ Run button.

Step 3: Set Up the API Key (10–15 min)

1.     Trainer explains: “To use the AI, we need an access key that tells the system who you are.”

2.     Demonstrate how to create an OpenAI account at https://platform.openai.com/signup.

3.     After login, click View API Keys → Create new secret key.

4.     Copy the key (a string of letters and numbers).

5.     Paste it into the second cell in Colab, replacing:

openai.api_key = "YOUR_API_KEY_HERE"

Trainer’s tip: For data safety, remind participants not to share keys publicly. If possible, provide your own key for class demonstration to avoid confusion.

Step 4: Run the Example Code (20–25 min)

1.     In the third code cell, participants see this sample:

import openai

# Connect to OpenAI

response = openai.ChatCompletion.create(

  model="gpt-3.5-turbo",

  messages=[

      {"role": "system", "content": "You are an educational assistant."},

      {"role": "user", "content": "Explain the water cycle for 10-year-old students."}

  ]

)

print(response['choices'][0]['message']['content'])

·  Trainer walks through each part:

·       model: tells which AI brain is used

·       messages: sets up conversation roles

·       content: the user’s prompt

·  Click ▶ Run.

·       Output appears below, showing AI’s text (e.g., “The water cycle is how water moves around the Earth...”).

·  Encourage participants to modify the user content (e.g., change topic or audience):

·       “Explain photosynthesis for teenagers.”

·       “Summarize the causes of World War I in three sentences.”

Step 5: Experiment and Reflect (20–25 min)

1.     Participants test new prompts, such as:

o   “Create 3 quiz questions about the human body.”

o   “Translate this paragraph into simple English.”

o   “Generate a fun classroom activity for learning fractions.”

2.     Trainer explains that this same technique can power chatbots, feedback tools, or quiz generators.

3.     Discussion:

o   What kinds of educational tasks could benefit from AI APIs?

o   What are the ethical or data protection concerns?

o   How can educators stay transparent about AI use?

4.     Optional extension: Participants add a new “system” message like:

{"role": "system", "content": "You are a quiz generator for language learners."}

Then run again to observe how output changes.

Step 6: Group Sharing (10–15 min)

·       Participants share one creative or useful example they generated.

·       Trainer collects ideas on a Padlet or shared document (e.g., “AI for feedback,” “AI for question generation”).

Wrap up by linking to further exploration (e.g., Hugging Face APIs, no-code AI tools like Poe.com, or Replit).

Adaptation Tips

On-site:Trainer demonstrates on projector while participants follow along on laptops. Trainer circulates to assist with copy-pasting code and troubleshooting errors.

Online:Trainer shares their screen in Zoom/Teams, provides the Colab link in chat, and supports participants through breakout rooms. Participants share results by copy-pasting outputs into chat or Padlet.

Hybrid:Trainer demonstrates in person while remote participants join via screen sharing and shared Colab notebook. All participants test and modify prompts collaboratively in real time.

Trainer Tips for Non-Technical Educators

·       Run the example yourself once before class to confirm it works.

·       Always use a shared demo API key (not personal ones from participants).

·       Reassure learners that copying and modifying existing code is part of the process — understanding comes gradually.

·       Focus on pedagogical implications, not coding mechanics.

·       Encourage exploration: one participant’s curiosity often inspires the group.