Instructional Technology

What is generative AI?

• Generative AI is a type of computer program that is trained so that it can create new content
• It can make text, images, music, and more
• Examples you may have heard of include ChatGPT, Google Gemini, Microsoft CoPilot, and the image generators DALL-E, and Midjourney

How does generative AI work?

• The large generative AI platforms like ChatGPT, Google Gemini and Microsoft CoPilot run underneath almost every tool that has an AI feature. They are called “Large Language Models” or LLMs.
• LLMs are trained initially by receiving very large amounts of data and content.
• Then, the training continues by instructing the LLM to analyze and recognize patterns and connections in the data it received.
• The LLM is then able to create new content based on what it “learned.”

Common uses of generative AI by the public

• Writing assistance (essays, stories, emails)
• Summarizing texts
• Analyzing data
• Getting step-by-step instructions for how to do something
• Image creation and editing
• Language translation
• Answering questions and providing information
• Generating music or art

Benefits of Generative AI

• Can help with contribute ideas and support brainstorming
• Can generate infinite numbers of examples, ideas or drafts very quickly
• Provides quick access to information
• Can assist with learning new topics
• Helps break down complex ideas and categorize information 

What are the risks, concerns and limitations of generative AI?

When using tools like ChatGPT, Grok, Google Gemini and Microsoft CoPilot, everyone should know that there are risks. Generative AI tools can include the following risks:

• Lack of data privacy and security, especially with free versions.
• Mistakes/errors
• Inventing “facts”
• Bias in the content generated
• “Echo chamber” effect and dangers of humanization
• Environmental impact: amount of energy needed to respond to each prompt plus the energy and/or clean or pure water used to cool the servers
• Negative impact on critical and creative thought if overused or misused.

So what do we do?

Use AI when you don't have the time, knowledge, and/or skill to do what you need, but don't use it all the time. Select the right tool for the task, and remember the risks, concerns and limitations. In addition, 

• Remember that YOU drive the work with AI. It depends on you in order to do its job well.
• The human is ALWAYS the responsible party. AI is NEVER responsible.
• Use AI thoughtfully and only when needed.
• Review everything AI generates for accuracy, tone, style, and bias.
• Talk to your students about AI and steer them towards only using tools their teachers are providing.

Resources  

• Parents’ Ultimate Guide to Generative AI from Common Sense Media
• KVIE "Abridged" online publication September 2025 article about generative AI in education the Sacramento region, featuring San Juan Unified.

 

 

 

Generative AI tools in San Juan Unified: Non-Negotiables for Use

• We only provide staff (and students) with generative AI tools that are safe for students
  • Data security and privacy aligned to California’s rigorous standards
  • “Closed” network: no prompts or generated content are shared outside of sanjuan.edu
• Tools are accessible to those with special needs
• AI tools “speak” all of major languages of our district families and employees
• Purchased tools are designed by teachers and/or receive special training in education to better support education-specific work
• Staff review, edit and revise all AI-generated content before using or sharing.

How staff use generative artificial intelligence

• Creating lesson plans and materials
• Generating practice and assessment activities for students
• Providing content that is responsive to students’ needs
  • Changing the reading level of a text
  • Explaining something in a way that a learner can understand or make connections to
• Translating content for students
• Drafting feedback on student work based on rubrics and instructions provided by the teacher

Our licensed tools that include AI components

• MagicSchool: a generative AI-powered tool with 80 templates for tasks educators do regularly, plus a chatbot to support staff with their work. Consider watching this video overview of Magic School, which was created by Magic School for educators.

• Read & Write, Kami and Pear Deck are tools that we have had since Distance Learning. They now have some AI-powered features.

  • Read & Write is a support tool bar staff and students can open at any time in their Chrome browser. The toolbar works with online content. Among the many tools are AI-powered features, such as the translator, text leveler, read aloud tools and more.

  • Kami allows staff and students to interact with a Google Doc or PDF in a number of ways. AI-powered features include translation, “explain,” text leveling and summarizing content.

  • Pear Deck allows teachers to add interactive content to their lessons that students receive on their Chromebooks. Some tools staff use to create this content are powered by AI. Students have no interaction with AI in this platform. 

• Canva and Adobe Firefly are multi-media tools that are available to staff as well as students in grades 6 and up to create content. This includes text-to-image tools that use AI to generate images.

The future of AI in education

• Teachers still design and direct learning experiences that meet the needs of your students and review all work before assigning a grade. They may use AI to support their work.
• AI will likely become more common in schools and elsewhere
• AI may support educators to offer personalized learning experiences
• Students will need to learn how to work ethically and productively alongside AI in order to function in an AI-integrated world.

 

Why provide a generative AI platform to students?

• Promote equity:
  • “Blocking” AI just worsens the opportunity and skills gaps that already exist between students who have consistent access to technology outside of school and those who do not. 
  • Students who have access to technology are already using AI tools (even if they don’t know what they should know about them) to get help and they are building valuable new skills related to AI use, such as prompting.
• Preparing students for a future that will include AI
  • Unlike adults, our students will never know a future without AI. 
  • There are new skills they (and all of us) need in order to live and work in an AI-integrated society, such as enhanced media literacy and prompt engineering skills, among others.
• Academic supports 
  • Safe, educational-trained AI tools can provide learners with valuable and needed academic supports any time they need them so that they can make progress towards the learning targets and their own personal goals.

Our licensed generative AI tool for grades 2 and up

Within MagicSchool, teachers in grades 2 and up access “MagicStudent” to create safe, structured, digital spaces with student-facing AI tools to

• Learn about AI
• Build new skills, such as prompting skills
• Support students’ academic achievement, critical thinking, and creative expression

Like the staff tool MagicSchool, MagicStudent has dozens of tools available, but students only access tools that are shared directly with them by their teachers.

Multiple training opportunities and resources are provided to teachers and other staff to support effective, ethical, and thoughtful use of generative AI, including programs specifically designed to help them with MagicStudent.

How was MagicStudent selected?

• First, it met all of our requirements for data privacy and security, accessibility, and language support for oral and text input and output in all of our major languages. 
• MagicStudent was selected through a pilot process where group of teachers in grades 2 and up received training to try MagicStudent with their students for a variety of purposes. Administrators and families whose students were involved also received information that the pilot was happening.
• Feedback was collected from teachers, students and families in the pilot group.
• In addition, the administrator who coordinated the pilot observed many classrooms while Magic Student was being used and shared observation data with the teachers.
• After the pilot, the teachers met to review all input and data received, and used that data to come to agreement on whether or not to recommend purchasing the license.  

Watch a video about MagicStudent.

How does Magic School Verify the Quality of Their Platform?

MagicSchool does regular testing and ongoing monitoring of the outputs from both the open-ended chatbot Raina and the dozens of tools in both the staff and student platforms.

They built in a “safety loop” that helps increase the quality of the content (“outputs”) generated. This loop has three parts:

• Framing user prompts responsibly before they reach one of the underlying models that will generate the output (response)
• Auditing outputs (responses) to check for accuracy, safety, and compliance
• Refining safeguards as models change and usage patterns shift
   

How will students access tools from MagicStudent?

Unlike other content students use, there is nowhere for a student to see a tile called "MagicStudent" and the more than 50 tools available in the platform.  This is intentional. It is not supposed to be something students access on their own, without the support of their teachers. Instead, teachers continue to make educational decisions to best meet their learners' needs by following the steps below when it comes to granting access tools to students:

1. Teachers review the available tools and even pre-built rooms and decide which tools to provide to students, customize those tools to better support their students and the targeted learning outcomes 
2. Teachers create “rooms” and put the specific tools they have chosen  in the rooms for their students
3. Students receive a link to these “rooms.” They never see an option to open “MagicStudent,” only the specific rooms their teachers created.

How do teachers monitor the use of the tools and the quality of the AI interactions with their students?

Teachers can see many kinds of data, including:

• All student interactions with AI in the rooms teachers create

• Every tool every student uses
• How often and when students engage with each tool
• Every word between their students and the AI tools the students are given access to

In addition, teachers will receive a message from MagicSchool (the staff platform) if a student is using a tool in ways that may indicate attempts to: 

• Circumvent safety features, 
• Cause harm to themselves or others, 
• Promote violence or hate, or create misinformation. 
• Teachers can follow up with students and involve other staff if necessary.

Will San Juan Unified Buy or Use an "AI Detector?"

There are companies marketing “detector” tools, claiming that will help teachers know if students are using AI unethically to create their assignments for them.

San Juan Unified will NOT buy “AI Detector” tools because 

• They have very low reliability and misidentify both ways: saying some used AI who didn’t and others did not use AI when they did.  Even ChatGPT stopped marketing the tool they created because it didn’t work.
• They disproportionately misidentify students living in poverty and English learners of having used AI when they did not. 
• Instead, we are supporting teachers to use equitable, standards-aligned instructional strategies, which will reduce the likelihood of students feeling unsure and being tempted to turn to AI
• We also provide staff with resources and training to teach students about AI and regularly have collaborative conversations with learners about what it means to have integrity regarding the work they turn in to their teachers (and elsewhere).

Resources

MagicStudent Frequently Asked Questions (FAQ) documents for students:

• Grades 3-5
• Grades 6-12

MagicStudent FAQ document for families

The Safety Loop for Students (whitepaper by Magic School explaining their safety protocols and procedures)

Getting started with conversations about generative AI

Ask them questions about their understanding of and experiences with Generative AI. Some questions might include:

• Has your class talked about generative AI?
• Have you used technology at school that includes AI?
• What do you think of using generative AI to support you with school work? 
• Are you excited to learn how to create with generative AI, for example, trying to create your own chatbot? What would you like your own chatbot to do?
• What kinds of things do you think you should not use AI for?

You can also share basic information about generative AI (see Part 1 of this series). Consider sharing responses to these questions:

• What is AI and how is it trained?
• What kinds of things can it do well?
• Why does it make mistakes?
• Why do the results reflect human biases?
• Why can AI tools be considered “insecure,” or unsafe and what does that mean for me?

As appropriate, consider and discuss deeper questions (see the rest of this part of the series)

• What is the role of the human when working with generative AI?
• What does it mean to use AI ethically?
• How can developers reduce bias in the results? How can users reduce both mistakes and bias by careful prompting?
• Why do I have to be careful not to “humanize” the AI tools and also about the “echo” effect of non-education-trained tools like ChatGPT and others?

Tools your students may have heard of but should NOT use

• ChatGPT, Google Gemini, Grok, and Microsoft CoPilot: Answers questions, writes text, creates images, and more. 

• DALL-E, Stable Diffusion, Leonardo and MidJourney: Create images/art from text descriptions supplied by the user
• Grammarly: Helps with writing and grammar

NOTE: if students are using Grammarly while signed in to their San Juan emails, this is safe to use. If they are not signed in, it is a different version of Grammarly that could sell or share their personal data.

Why shouldn't students use tools not provided by San Juan Unified?

• Some may produce/share inappropriate content with students
• Most public tools require accounts with personal details: They might collect, sell and share personal information, including the work students do in AI. This data might be used for advertising, training, or even generating other people’s content. In addition, it is far more likely to be included in data leaks.
• Tools like ChatGPT may provide harmful suggestions (mental health, for example) and create harmful content if students feed them misinformation, violent content, etc. 
• AI can sometimes give wrong or biased information, especially when not trained for educational use
• AI tools like ChatGPT will do work for students rather than being a support or guide

Ethical considerations, especially with tools like ChatGPT

• AI-generated content might be used without permission or in unethical ways, for example, students may use AI in place of doing their own work
• Students (and all people) may over-rely on AI tools, leading to loss of critical thinking skills and creative capacity
• AI tools may use the copyrighted works of others (art, literature, film, images, music, etc.) without the owner’s permission or even credit to the owner when generating content
• AI could be used to create misleading and harmful content

AI issues: accuracy and reliability

Using tools that have not been trained for use by staff and students in education settings are at greater risk for mistakes and other issues

• AI can make mistakes, make things up or provide outdated information
• It may not always understand context or nuance
• Results can be biased based on the data used to train the AI

Remember that the human must always double-check information provided by AI tools

Supporting your child's learning with AI

• Encourage using only San Juan Unified-approved resources 
• Help your child understand when to use AI and when not to
• Promote creativity and original thinking alongside AI use
• Remind students that they are responsible for their work and their learning! The AI is NEVER responsible.
• Share how you use AI ethically to support you; talk about how you evaluated the content, what you decided to next with that content and why.

Engage in conversation with your student

Consider asking them questions about their work, such as the ones below:

• What were the instructions?
• What was your favorite part of the assignment? Why?
• What was difficult? Why? How did you move forward?
• What technology tools did you use? Were they recommended by your teacher? How did you use each tool to support you? 
• What are you most proud of in your work? What do you hope your teacher notices? 

Protect your child online

When students begin using the internet in general, you can do these things to support them:

• Supervise your child's internet use
• Use parental controls when possible
• Teach children about online safety and privacy
• Encourage open communication about online activities
• Don't share personal information with any tech tools
• Be cautious about believing everything AI tells you
• Use AI as a helper, not a replacement for learning
• Always review AI-generated content before using it

Resources

Parent's ultimate guide to generative AI from Common Sense Media