Leon Furze

This is the third post in a series exploring how generative artificial intelligence tools can be a part of the writing cycle to help students improve their writing technique, style, and structure. For the previous post on the ‘exploration’ stage, click here.

Idea generation can of course happen at any time during the writing process, with inspiration striking at the most unexpected moments. Sometimes, I start writing without a clear idea in mind, only discovering what I’m writing about halfway through. But in the idea stage in the writing cycle, students get the chance to deliberately deepen their understanding of the purpose, audience, and context for their writing after they’ve looked at exemplary works through mentor texts. Importantly, this stage of the process aims to help with the common fear of staring at a blank page, a challenge faced by all writers, from students to seasoned authors.

There’s a great post from The Marginalian discussing the challenges writers face with the blank page and creative block. It features insights from eight celebrated writers, including Joyce Carol Oates, Margaret Atwood, and Jonathan Franzen, who share their perspectives on overcoming the paralysis that can come from facing a blank page. All of the authors highlight the importance of action and ideation, with strategies like accepting bad writing as a warm-up exercise and understanding that a blank page is also a door to infinite possibilities – strategies that I’ll touch on here, and in a later article devoted to first drafts.

Join the mailing list to stay up to date on future articles:

Some view Generative AI as a threat to “traditional” writing methods, while others see it as a potential cure for the fear of the blank page. For instance, Microsoft Word’s Copilot feature (accessible through Copilot for Microsoft 265, or Copilot Pro) allows for the automatic generation of entire drafts with a prompt directly within the word processor.

In the next couple of weeks, I’ll write a longer post focusing on why relying on AI to draft entire pieces might not be the best approach. Nonetheless, AI can serve as a tool for supporting students’ own knowledge, creativity, and ideas, especially in brainstorming and idea generation.

Practical AI Strategies includes an entire section on GenAI policy and assessment. It is available from Amba Press

Activities for Ideas

Activity one: Context Brainstorms

Using generative AI for brainstorming offers an easy route to generating a list of ideas, though typically the output is generic due to the nature of large language models. This becomes obvious when multiple students use a basic prompt, like “brainstorm a list of ideas for a blog post about the climate crisis.” Despite the “stochastic” nature of language models which introduces randomness, results often show significant overlap. Instead, students need to use more contextualised prompts in order to generate more unique and interesting responses.

1. Discuss the importance of context: Start by discussing the importance of tailoring prompts to better reflect the specific project or topic students are working on. This helps produce more unique and varied brainstorming outputs. Refer back to the work on context for ideas about how a prompt might be contextualised.
2. Prompt variations: Explore how to adapt prompts based on the writing project’s nature. For example, prompts for nonfiction pieces like blog posts might differ from those for fictional short stories, involving variations based on writing style, subject, or audience.
3. Generative AI prompts: Encourage students to apply these more contextualised prompts in their brainstorming sessions, using tools like Microsoft Copilot. Use the table below to offer a few ideas, and look at the varied examples from Copilot, ChatGPT and Google Gemini.

Table of context terms to use in brainstorming prompts

Category	Terms
Tone	Informative, Persuasive, Descriptive, Reflective, Sarcastic, Humorous
Style	Formal, Informal, Academic, Conversational, Narrative, Expository
Form	Essay, Blog Post, Short Story, Report, Poem, Article
Perspective	First-person, Second-person, Third-person
Audience	Peers, General Public, Experts, Children, Teenagers, Adults
Purpose	To inform, To persuade, To entertain, To explain, To argue
Subject Matter	Environmental, Technological, Cultural, Historical, Personal, Political

Examples

Activity 2: GenAI Thinking Routines

As I’ve said elsewhere, I’m a big fan of the Project Zero Thinking Routines, and they offer a great way to extend students’ brainstorming with GenAI. For example, students could do their own brainstorming individually or in groups, and then take their initial ideas and run them through a routine like Generate Sort Connect Elaborate with a prompt like the following, in an example taken from a Design and Technologies subject:

These are our group’s ideas for a sustainable furniture item that will respond to this client’s brief <copy/paste design brief>: <copy/paste initial brainstorm>. Use the Generate Sort Connect Elaborate thinking routine to help us extend our initial ideas – go through the routine like you are a member of our group

Basically, students complete the “generate” stage of the Thinking Routine, and then hand over to GenAI to add a few suggestions and assist with the S-C-E steps.

1. Show students the basic structure of the Generate Sort Connect Elaborate thinking routine on the Project Zero website
2. Have students complete the first step, “Generate” individually or in groups
3. Use a prompt like a variation on the one above to sort, connect, and elaborate upon those ideas.

Example from Copilot

Activity Three: Brainstorming with image gen

AI’s ability to generate images quickly and without the need for endless stock library scrolling (or Google image search, another classic student time-sink) presents an interesting brainstorming method, offering visual inspiration that can drive written content. This approach is especially useful for abstract concepts or when seeking to add a visual element to digital texts.

1. Choose an image generation platform: Select a platform that students can access, like DALL-E (in Copilot) or Adobe Firefly (which has a K-12 license), allowing for image creation from textual prompts.
2. Initial free writing session: Have students engage in brief free writing to brainstorm as many ideas as possible within a set time. This could be as prose, dot points, mind map, by hand, or electronic. You could also combine the previous two activities into this stage.
3. Create image prompts: Use the free writing output to form image prompts for the selected platform. Using a chatbot like Copilot, students could copy their entire free writing as the prompt and tell the app to generate an image. In Firefly, students would need to either turn their free writing into an image prompt, or generate an image prompt based on the writing in another chatbot.
4. Discussion: Reflect on how the generated images can enhance the brainstorming process and aid in developing ideas for written content.

Activity 4: Idea Incubator

This activity is adapted from Practical Writing Strategies to incorporate GenAI. The idea incubator is about finding time and space for brainstorming. Often, we’re hindered by the time constraints imposed on us, and perhaps even the location we’re working in – that’s especially true when teaching writing, where the emphasis is on getting to the end-point of an assessment piece. This strategy encourages teachers and students to dedicate time to brainstorming and finding space to do it.

1. Set aside dedicated class time for brainstorming. A change of setting may also help. Consider sending students outside (weather permitting) or to the library.
2. Ask students to write down any ideas that come to mind, no matter absurd they may seem. Allow students the time and space to wander in their thoughts.
3. Once students have a list of ideas, ask them to take a step back and look for common themes and connections. This will help them to identify the core concepts they’ve written.
4. For each core concept, students write a sentence that summarises the idea and explains what makes it interesting or unique. This will help students develop a clear focus for their writing.
5. Take these summary sentences, and pose them as questions or statements in a Generative AI prompt. See what the application comes up with, and whether it extends those initial ideas or adds any interesting angles.

Conclusion

Hopefully you can see from these activities that using generative doesn’t need to entirely replace students own idea generation or brainstorming. In fact, using GenAI straight out of the gate as an idea generation tool usually isn’t a good idea. There is research emerging that as a creative assistant, generative AI is only really helpful for people who would ordinarily score very low on creativity indexes. For everyone else, it’s average at best.

Contextualising the prompts using your own ideas as the basis and thinking of ways to use these technologies instead of generating the entire first draft is much better. The next post in this series will focus on how generative AI can be used to support the explicit instruction of specific writing skills and techniques.

If you have questions or comments, or you’d like to get in touch to discuss GenAI consulting and professional development, use the form below: