Transforming Worlds Into Words: Using ChatGPT to Bring Student Visions to Life

Abstract

This action research study evaluated an original lesson using the AI language model ChatGPT to generate tailored play scripts for individual students based on their interests, reading levels, and creative visions. Implemented with 13 third and fourth graders in an after-school program, the “Journeying between Worlds and Words” lesson aimed to empower students to translate their imaginative worlds into written scenes they could perform. The ChatGPT-generated scripts promoted high engagement, with students excited to see their ideas brought to life. The scripts provided authentic opportunities to build reading fluency through repeated readings and comprehension by dissecting characters, settings, and events. Students demonstrated creativity in elaborating on storylines and props. Using ChatGPT proved feasible and efficient for obtaining appropriately leveled, high-interest scripts compared to pre-existing scene collections. The lesson showed promise for differentiating instruction and empowering student voice and serves as a model for other projects. Rather than struggling to find the right book/script, teachers could harness AI to generate relevant stories matching each student’s needs and passions. Overall, this lesson highlights AI’s potential for transforming students from consumers of learning experiences to creators of their own.

Keywords

differentiation instructional strategies talent development the arts artificial intelligence twice-exceptional technology

“ChatGPT enabled us to create the ideal engaging, differentiated scripts to spark students’ interest and fuel their motivation in the performing arts curriculum.”

“What’s the world for you if you can’t make it up the way you want it?” - Toni Morrison (1992)

The students were beaming when they received their playscripts. They asked for copies to share with their family and classmates because these scripts were special. These scripts translated the students’ own worlds into printed words, into a playscript they could read with their friends. This experience was part of “Brilliant Storytellers,” a new third and fourth grade curriculum designed with a dual purpose: to enhance reading skills, foster creativity, and develop leadership abilities through performing arts, while also serving as a mechanism to identify leadership and creative talent in young learners. Over the past two years, we have been crafting lessons that use theater as a vehicle to engage students in reading fluency, strategies, and comprehension skills. By inspiring students to bring their stories to life on stage, we’ve created a powerful motivator for them to improve their literacy skills while simultaneously nurturing and assessing their creative expression and leadership potential.

We started this journey with some naiveté and a lot of optimism that this would be a straightforward curriculum to develop. However, we quickly realized several significant challenges, including the most perplexing: finding playscripts. Our curricular vision not only required extraordinary scripts, but a lot of them. We needed scripts students wanted to read and rehearse, enough scripts for every student to direct one. Once in a pilot lesson, we gave students an existing script, and we heard an audible groan. They quickly reminded us that our club was a part of an afterschool program, not a reading class.

Part of the challenge was our myriad criteria for the scripts. We were asking for the moon thirteen times, and that was just for our current pilot study. We needed scripts that were relatively short (2–3 minutes in length) and matched students’ reading ability levels while incorporating age-appropriate content. Further, scripts needed to be incredibly engaging for third and fourth graders, as our program is implemented after school and in the summer. We wanted the scripts to connect to student interests. As these interests were quite varied and changed rapidly, adult authors would struggle guess youth’s interests, and we needed a way get their actual interests into the scripts.

Our challenges reflect the challenges many classroom teachers face. For example, third-grade teachers are often tasked with serving students across a significant range of reading abilities and needs in a single classroom (e.g., students may span 9.2 grade-levels in reading comprehension). Further, these students tend to develop reading skills at different rates, so over time, the range only increases (Firmender et al., 2013). Further, much research has demonstrated teachers struggle to find the time, resources, and support necessary to differentiate the curriculum effectively (VanTassel-Baska et al., 2021; Ziernwald, 2022). We understand it is a daunting task. Ideally, we would have scripts representing students’ passions or exposing them to new ones; however, even if these perfect scripts exist, it requires the herculean effort to curate a classroom library and match students with appropriate texts.

Toni Morrison (1981), as cited in Brown, (1981) provided insight on how to address this issue: “If you find a book you really want to read, but it hasn’t been written yet, then you must write it.” This sounds straightforward until you are faced with a blank screen or sheet of paper. Even as adults communicating our vision clearly with words seems daunting. Many elementary and middle school students find it almost impossible to translate the worlds they see in their minds into words. This experience is an example of asynchronous development (i.e., the mismatch between cognitive, physical, and emotional development; National Association for Gifted Children, 2024). Stephanie Tolan (1994/2024) gave a perfect example of this:

A gifted child may appear to be many ages at once. He may be eight (his chronological age) when riding a bicycle, twelve when playing chess, fifteen when studying algebra, ten when collecting fossils and two when asked to share his chocolate chip cookie with his sister.

We see these asynchronicities with these students. Their ideas, their worlds are intricate and represent the stories they want to read, yet even if students could write their worlds, it would require additional time and effort to convert these ideas into script format, write stage directions, and develop a dialogue-focused work as opposed to a narrative one. Elementary students often have more experience reading and writing narrative paragraphs as compared to reading and writing playscripts, so building their worlds as a playscript presents an additional challenge. If we and our students cannot communicate our ideas fully in the time frame and we cannot find existing scripts that fit our requirements, that puts our entire secret reading mission in jeopardy.

We needed an informant/asset, which is why we turned to generative AI. We decided to flip the script (pun fully intended) and use generative AI to generate scripts tailored to each student’s readiness level and interests. We sought to develop a transferable pedagogical approach to address these issues. In this article, we present our experiences with the implementation of the lesson, “The Journey between Worlds and Words.” We provide a glimpse into the trials, tribulations, and triumphs of implementing our program and attempting to create custom-made educational content. We hope all teachers can use what we have learned and tailor it to their own classrooms.

Methods

This action research project evaluates the “Journeying between Worlds and Words” lesson, a component of the larger federally funded initiative, Project Brilliance, supported by the Jacob K. Javits Gifted and Talented Students Education Program. A key objective of Project Brilliance is to develop a comprehensive talent identification system for gifted and talented students, especially in traditionally overlooked domains. Our curriculum functions as a dynamic assessment tool, where students are scaffolded through tasks while being purposefully observed and evaluated. Each lesson incorporates structured talent spotting opportunities, allowing instructors to formally identify creative and leadership gifts.

This approach enables us to observe students’ potential in real-time, authentic learning contexts. To complement and validate our observational talent spotting method, we are concurrently developing task-based assessments. These additional measures aim to provide further evidence of the effectiveness and accuracy of our talent identification approach, ensuring a robust and multi-faceted system for recognizing gifted and talented students. Preliminary data suggest our approaches have identified both overlooked creators and leaders, specifically when comparing that of teacher rating scales in creativity and leadership. These findings will be presented in future work after being replicated across multiple cohorts.

Participants

All third and fourth grade students (n = 17) in a pre-existing community afterschool program were invited to participate, and 13 of those students were present to participate in this lesson series. The students’ reading levels mimicked a traditional mixed ability classroom, with students ranging in reading levels from kindergarten to fifth grade. These students attended a k-4 elementary school in a Midwestern, suburban area, which served 239 students. Within the school, 76% of students are white, 5.4% Black/African American, 4.6% Hispanic, and 12.1% multiracial. Further, about 71% of the students were identified as economically disadvantaged. The instructors included two university professors and three undergraduate education majors.

Materials and Procedure

The broader curriculum, Brilliant Storytellers, includes modules and lessons dedicated to acting, directing, and designing to support students developing reading, creativity, leadership, and performing skills. However, the purpose of this case study is to examine one lesson, “Journeying between Worlds and Words.” Embedded early in the directing module, this lesson provides the opportunity for students to develop their vision for a scene they will direct. We purposefully selected this lesson for this article because it presents a potential solution (i.e., Generative AI like ChatGPT) to a consistent, ubiquitous issue—identifying appropriate, engaging reading options for a wide range of readers.

Lesson Objectives and Standards

This lesson spanned four days (1 hour per day), but the scripts developed in this lesson were used in subsequent lessons, as students directed their envisioned script. This 4-hour lesson was anchored on several key standards/objectives:

• Reading (CCSS.RL.3.3): Describe characters in a story (e.g., traits, motivations, feelings) and explain how their actions contribute to the sequence of events.

• Performing Arts and Theatre (TH: CR.1.3 A): Create roles, imagined worlds, and improvised stories in a drama/theater work.

• Leadership: Develop and communicate a vision for your team.

• Creativity: Generate many, different, original ideas to tell a story. Use constraints to build new ideas.

Importantly, writing was not the objective of the lesson. This lesson was not designed to teach students how to write a script but rather, the script was the tool to address the other lesson objectives.

Assessments

Assessments serve as the bridge between the learning objectives and the actual lesson; therefore, we considered and implemented several types of assessments to track growth towards the objectives. To create effective lessons, all students needed to be pre-assessed, especially in reading to best develop the AI prompts. In a traditional classroom environment, this could be done using Dibels and/or other reading fluency/comprehension measures. Within our program, however, we used observations while students were cold reading scripts. We also used fun pre-planned lessons to gather informal classroom observations data on students’ abilities in leadership, creativity and performance arts skills.

For example, to assess students’ leadership levels, we played a game, Home Helpers, during which students took turns leading small groups. The groups were tasked with working together to become a common kitchen appliance, and then, the other teams guessed which appliance they were. During that game, instructors observed when each student leader for the development of an idea and then, clear communication/direction making that idea happen. Then, we could use that information to (a) build groups that paired confident directors with nervous actors or (b) provide additional scaffolding for those that struggled. Each objective had aligned pre- and post-assessment options, which are presented in Table 1.

Table 1.

Pre- and Post-Assessments for the Lesson Objectives.

Objective	Assessments	Exemplary Criteria
Reading Describe characters in a story (e.g., traits, motivations, feelings) and explain how their actions contribute to the sequence of events	Pre-Assessment: Previous playscripts’* character analysis using the GOTE worksheet (see Figure 3). Previous playscripts included Sideways Stories from Wayside High and Andre the Chef	• Clearly identifies the characters’ main goals, obstacles, tactics, and expectations with detailed descriptions • Explains how the characters’ goals obstacles, tactics, and expectations influence the story’s events and outcomes
	Post-Assessment: Personalized playscripts’ character analysis using the GOTE worksheet (see Figure 3)
Performing Arts and Theatre Create roles, imagined worlds, and improvised stories in a drama/theater work	Pre-Assessment: Tableau Environments—Groups are given a biome (e.g., desert), and students need to silently contribute to the setting (e.g., becoming a cactus)	• Develops a detailed, imaginative, and coherent world that enhances the story (e.g., uses props, uses vivid descriptions in the reflection) • Demonstrates originality in the setting, characters, and/or plot components • Describes why they made certain decisions and how they may improve upon this work in the future
	Post-Assessment: Final taped performance and student reflection
Leadership Develop and communicate a vision for your team	Pre-Assessment: Home Helpers—students take turns as the leader, and they need to direct their group to become an assigned kitchen appliance (e.g., a toaster). The other students must guess what they are	• Presents a clear and detailed vision for the desired outcome • Communicates the vision in a way that others understand and can follow • Engages team members in pursuing the vision
Leadership Develop and communicate a vision for your team	Post-Assessment: Team observations and student journal pages (GOTE and blocking pages)
Creativity Generate many, different, original ideas to tell a story. Use constraints to build new ideas	Pre-Assessment: This is Not a Pen Activity—the instructor holds up a pen and states, “This is not a pen; it is a toothbrush.” The leader passes the pen to the next student to develop as many different uses for a pen. New items could be introduced, like a stapler or whisk	• Generates multiple possible responses, characters, settings, and/or plot points • Develops an original vision for their story that includes unique responses, characters, interesting settings, and/or intriguing plot points • Uses the materials and space available to communicate the story
	Post-Assessment: Observations of small group discussions and final taped performance

Procedure of Broad Learning Experiences

The curriculum’s objectives and corresponding assessments guided the development of our learning experiences. While we maintained a consistent structure for all students, we tailored the content and level of support based on pre-assessment information. Our consistent structure was as follows:

• Opening Activity: We began each session with a transitional activity to help students shift from their school day mindset to the afterschool club environment. For example, in one activity, students would silently act out their favorite food for others to guess.

• Introduction of Genres: Then, within this lesson, we began by watching four different movie trailers from each of the four genre options. Although there are more than four genres, we limited it to fantasy, adventure, mystery, and science fiction due to time constraints. For example, for the adventure genre, we showed the movie trailer for Honey, I Shrunk the Kids. Collectively these trailers were perfect short glimpses into the story that allowed all students, independent of whether they watched the full movie, to participate in an analysis and discussion of the genre.

• Exploring Multiple Worlds: Then, we set up four centers around the room to discuss each genre. The students brainstormed the genre’s key features, including character and setting possibilities. They visualized these worlds in their journals.

• Selecting One World: After considering each genre, characters, and settings, they selected the one that they would want to direct. Then they completed several prompts to further delineate their vision.

• [Instructor Work—Behind the Scenes]: After this session, the facilitator entered each student’s responses (character names, genre, and setting) into a generic ChatGPT prompt, which would then generate a script that aligned with the students’ vision.

○ The ChatGPT prompt was “Please create a script for two characters at the [third grade reading level] within the [genre type] genre. It should include the following characters: [character 1 description] and [character 2 description]. It should take place in [this setting]. Also, it should [if students provided any additional requests]. Scripts should take 3 minutes to read aloud.”

○ We copied and pasted the prompt, filled in the unique pieces, and within 2 hours, we had 13 scripts.

• Sharing your World: Students were placed in groups of three to experience their world’s translation into words. Each student director had the opportunity to observe two student actors read the script. Then, they highlighted key words that needed defined and analyzed the main characters’ goals, obstacles, and tactics. They discussed and incorporated any necessary edits. In their scripts, the students also described potential vocal variation options (e.g., emphasis, pitch, rate, and volume).

• Blocking your World: Finally, students planned the directing of scenes by documenting their ideas for the scene on paper. Specifically, using a 3 by 3 grid representing the upstage, downstage, center stage, stages right and left, they applied blocking notation to plan the actor movements. Each acting unit or beat could be contained within one blocking grid and would correspond to a particular section of text. Pre-blocking the show allowed the students to communicate with their peers more efficiently about their ideas. It also allowed the student directors to stay on track. As their work time ticked down, they could look at their blocking maps and clearly see how much more work they needed to do. When the actors had questions, the directors could show them the blocking patterns.

After this preparation work, in subsequent lessons, students led rehearsals with their actors, implemented blocking patterns, gave notes, and brought the script to life. Finally, the instructors filmed their rehearsed scene to show in a film festival at the end of the module.

Additional Opportunities for Increasing Depth and Complexity

Throughout this lesson, there are multiple opportunities to increase the depth and complexity of instruction for gifted learners. We integrated many of the performing arts/leadership/creativity opportunities on subsequent days because this single lesson was already packed into four days, but the students really benefited from these ideas in subsequent sessions. The reading differentiation happened within the current lesson.

Reading

If the pre-assessment demonstrates students can clearly identify characters’ GOTEs, then, additional abstraction and complexity can be added. For example, the initial prompt for ChatGPT to include higher reading levels, incorporate deeper themes (e.g., “incorporate the theme of change”), or build more complex relationships (e.g., “incorporate a relationship that gets weaker over time”). These edits will make the GOTE practice more complex and nuanced.

Performing Arts and Theatre

If students demonstrated their ability to build imaginative worlds in the pre-assessment, a way to challenge them is during their self-reflections. Ask students to watch their final production and critique their world building. Provide criteria or encourage students to develop their own (e.g., characters’ voice, movement, staging, props). After they have reflected upon their own, show a short video from the Lion King on Broadway Educational Series: Episode 1. In this 6-min episode, they discuss how they turned an animated movie into a stage production. Students can then map out what they imagine/sketch/describe what they could do differently for their own work that would make it more dynamic. This could also include adding new blocking patterns to better communicate their story, add pantomime or actor expressions to deepen character elements, and utilize the whole performance space.

Leadership

To continue to challenge students who have strong visions, provide authentic director training by assigning a chapter from the book, Directing Actors (Weston, 2021). This book was written for aspiring professional directors and examines the subtlety and multidimensionality associated with establishing a vision and supporting the cast.

Creativity

For creatively advanced students, they should be placed in the role a practicing professional. One potential extension would be to consider costume designing. Ruth E. Carter is the award-winning costume designer for films such as Black Panther and Selma. She has a 6-min Creative Spark video during which she discusses her creative process when designing costumes. Encourage students to try out her strategies for their own work, to consider how different costumes tell different stories.

Data Collection/Analysis

We collected several different pieces of data to evaluate the feasibility/validity of this lesson. We video recorded the lesson, collected student journal responses, reviewed their scripts, and had our own debriefing focus groups. For this article, we examined each of these pieces of data through an evaluation lens to identify moments that promoted student growth and engagement as well as moments that did not go as planned, which inspired additional curriculum changes. Throughout this process, we incorporated several checks of credibility and trustworthiness, including collecting multiple data sources across participants and ensuring multiple instructors read, revised, and confirmed the findings.

Findings

First, we present one example of a student’s experience within the script generation process, and then, we more broadly evaluate the success of this lesson by considering its (a) feasibility, (b) development of reading engagement, and (c) promotion of creativity. We could have examined a variety of outcomes, but these seemed to be the most prominent in our findings.

Oscar’s Midnight Adventures

At the end of the directing unit, we asked all students to rate the unit from 1–10 using their fingers. Oscar (participant’s pseudonym) quickly made two circles with his hands and pushed them together. He explained, “I loved it infinity. When you were acting, you can see all the possibilities you can do when you grow up, and directing was just amazing because you get to, like, tell stories and make them.” To reach that evaluation, Oscar developed his own adventure story, Oscar’s Midnight Adventures. Out of four genre possibilities (i.e., adventure, science fiction, fantasy, and mystery), he joined the adventure team. Within the adventure group, they discussed characteristics of the adventure genre (e.g., protagonists’ quests/adventures, danger, and/or exotic locations). Following that general conversation, students needed to select a setting. Oscar thought of a fire arena at midnight, two characters (Oscar and William), and several plot points (see Figure 1 for his unit journals).

Figure 1.

Oscar’s unit journal.

Oscar had a clear vision for the story he wanted to tell, but some students struggled. So, we developed a series of “Starting Sparks,” which are ideas for settings, characters, or plot points. Depending on instructor preference, they may be used only with struggling students or with everyone. See three “Starting Spark” setting examples in Figure 2.

Figure 2.

Starting Spark examples.

Then, we integrated his thoughts into the blank prompt and submitted it to ChatGPT3.5:

Please write a 3-minute play script at the 3rd grade reading level with only two characters. The play should be an adventure. The first character is Oscar Panszi, 21 years old, tall, often confused, and really funny. The second character is William Eliott, 24 years old, tall, serious, and powerful. The setting is at midnight at a fire arena. The plot should include the following events: Oscar strangely appears in an arena. He goes through various challenges and threats from William.

Here is an excerpt script from the first challenge:

William: You’re in the fire arena, Oscar. A place of challenge and danger.

Oscar: [Startled] Oh, hey there! I didn’t see you lurking in the shadows. What’s with all the fire?

William: It’s part of the challenges you must face here. Are you ready to prove yourself?

Oscar: Um, sure! But what kind of challenges are we talking about? [Looks nervously at the flames]

William: You’ll soon find out. [Raises his hand, and suddenly, fiery obstacles appear in the arena]

Oscar: [Eyes widen] Whoa! That’s intense!

[Oscar cautiously navigates through the obstacles, stumbling and tripping along the way. William watches with a stern expression.]

William: Focus, Oscar. You must overcome each challenge with determination.

Oscar: [Nervously] Right, focus. Got it!

[Attempts to dodge a flame, narrowly avoiding it]

Oscar: [Breathless] Phew! That was intense! I made it though, right?

William: [Nods] Impressive, Oscar. But the next challenge awaits. [Points to a massive flaming pit in the center of the arena]

With a script in hand, students were now ready for the last component within this lesson which involved preparing students to direct their scripts, which required bringing their scripts to life. We scaffolded students’ efforts by leading small group discussions of each script. Together we analyzed the characters, using the GOTE method. (See Figure 3). Developed by Robert Cohen (2002), this theater analysis tool provides an acronym to help directors/performers examine a character’s Goals, Obstacles, Tactics, and Expectations. This provides a clear framework for young students to use to better understand their characters. All instructors also helped scribe the students’ discussions, so all students could be a part of the conversation, independent of their writing abilities. In the future, students could also record conversations, and otter.ai would be able to provide a transcript. Depending upon the goal of the lesson, AI provides so many options to help students stay focused on the primary objectives.

Figure 3.

The GOTE worksheet.

After the character analyses, we discussed blocking and movements. We played a game where they learned different stage locations, and then, they worked to imagine how their two characters would move in their script. Figure 4 also includes Oscar’s blocking. After the script generation/planning process, each student director collaborated with two peers, who served as actors to vivify their theatrical piece.

Figure 4.

Blocking worksheet.

The students’ journey began with their own visionary premises, which ChatGPT then translated into a scripted story. The student directors were then able to use the script to communicate with their actors to bring breathe vibrant life into their once delicate, fleeting imaginings. This empowered students, like Oscar, to truly tell their stories.

Practically Feasible

This lesson was only possible because of ChatGPT. Curating an anthology of appropriately leveled and engaging scripts was an immense challenge both for us and teachers engaged in similar lesson planning. ChatGPT proved to be valuable in our curriculum development process. We had spent months reviewing existing scripts, struggling to identify selections that could meet all our criteria across the wide range of student reading abilities and interests. Within 2 hours, we had a full set of custom-made scripts, each one tailored to our students’ reading levels and interests (e.g., manga, peanut butter, cats, and “your mom”). Utilizing ChatGPT’s capabilities proved exponentially more efficient than the arduous task of scouring pre-existing materials to cobble together something suitable for our diverse learners. ChatGPT enabled us to create the ideal engaging, differentiated scripts to spark students’ interest and fuel their motivation in the performing arts curriculum.

Despite how awesome these scripts were, we needed to address some additional challenges. First, we needed to share these scripts with the students. Our first idea was to upload them to a cloud folder and let the students download them onto iPads; paradoxically, this technology was not working for us due to Internet restrictions. We decided to simply print multiple copies of each script, but then, several students requested edits. We were continually reprinting scripts to match their ideas, which was worth it to communicate how important their ideas are, but it affected scale-up feasibility. So, we had to make a choice. Either the kids handwrite their edits on the printed scripts and take responsibility for communicating their revisions to their team, or they learn to love the script as is. It’s not a perfect solution, but it’s one that’ll make this lesson more feasible in the future.

Our next challenge was part of a larger debate on the use of AI in the classroom; we did not use the software in front of the students, as we did not feel comfortable teaching students to use ChatGPT without significant conversations on AI ethics (Besnoy, 2023). While our program currently does not teach writing skills, we had mixed feelings about teaching them a potential “short-cut” to written works. We simply acknowledged the computer helped us write their scripts, but as an afterschool provider, we decided additional details or uses would be best left to the classroom teachers and caregivers/parents. For more information on how we would do that, see our discussion in the limitations section. Overall, with several future adjustments, we believe this is a practically feasible lesson in most settings.

Enhancing Reading Fluency, Comprehension, and Engagement

This work was worth it when we handed out their personalized scripts, when they went back and told their friends they should join Brilliant Storytellers when they are old enough, when they wanted to take them home with them. This excitement influenced participation in the repeated readings of the scripts. Often, students complain when asked to read the same passage more than once; however, rereading is an authentic component of theater. Actors rehearse multiple times, each time making different voice decisions or using different gestures. Each time is fresh, yet the words become more and more familiar, supporting students’ fluency development.

In addition to fluency, we could see students develop a deeper understanding and comprehension of their scripts. For example, each student director simply listened to their generated script being read. This freed the director to take notes on the script and of the performances. To support this process, the students circled words they did not know to increase student vocabulary, and they had a careful discussion to dissect each character’s goals and obstacles (See Figure 3). With that information, they read the script again making different vocal decisions or using different gestures because now after that dissection, they understood the vocabulary and their characters’ motivation.

We saw the most growth, however, in the level of student engagement. When we first handed out other scripts, there were many moans and complaints that they shouldn’t have to read after school. When we passed out their own scripts, they were elated. We saw further evidence of this when we visited them outside of our programmed time because we needed to make several edits to a script for a student. We worked with this student outside of our traditional program hours in the common cafeteria, meanwhile multiple other students heard we were developing personalized scripts. They came over to ask if they too could develop their own script. Students wanted to see their ideas come to life, to translate their internal worlds into words. And this lesson is one way to make that happen.

Students also became invested in ways we did not predict. For example, some students named their characters after themselves, and then, gave those characters some interesting characteristics. As described above, Oscar, the character, was really tall, often confused, and really funny. For Oscar, the student, two of these characteristics are aspirational at best. Yet, this could be a way to support students seeing themselves as the person they want to be. This also provides students the opportunity to place people like themselves in leading roles. For example, one female student included two female characters: a brave knight and the daring superhero. Using ChatGPT lowered the activation energy to bring to life the characters they want to read about and become.

Facilitating Student Creativity

This lesson also promoted creative thinking in ways we did not predict. We had imagined that students would demonstrate creativity as they developed their initial worlds, yet that lesson component had mixed results. Students needed to provide characters, settings, and potential plot points. Initially, we did not see much fluency or originality of thought. Even students who demonstrated creative thought in our pre-assessment did not translate that into their story line. Many students often included themselves and/or their friends as main characters. Further, to select the setting, many students needed some additional scaffolding, and many relied on our Starting Sparks.

Despite students not developing many or original initial ideas, students demonstrated a high degree of creative flexibility and elaboration in the final product. Even though their characters were named after themselves and their friends, students gave these characters interesting features like being 21 years old, teleportation powers, or the ability to shape shift. They added balconies inside their mansions and bridges in the woods.

Further, students took advantage of other openings to be creative that were embedded within their scripts. First, the software program generated scripts that had theatrically unrealistic stage directions for most school productions. For example, “suddenly, fiery obstacles appear in the arena” works well in a narrative text but are problematic in a play script. This led to the director needing to establish where the flames are coming out of the ground, marking those places, and working with the actors on how to avoid the flames. The challenge, though, was that while the director could see the flames in his head, the actors struggled to see past the X on the floor. Their reactions represented their developmental stages as they were in varying levels of abstraction.

Second, whether it is because of the underlying algorithm in ChatGPT3.5 or because we requested the scripts be no longer than 3 minutes, the storylines were sometimes quite generic. For example, in Oscar’s adventure script, ChatGPT wrote: “Oscar cautiously navigates through the obstacles, stumbling and tripping along the way.” This generic description provides an opportunity for the director and actor to work together to fill in the gaps: What are the obstacles? What does he trip on?

As another example, ChatGPT provided the following: “They begin rehearsing for the play, with Shama showing off her shape-shifting powers and Papa Frita trying to steal the spotlight.” Again, this vagueness provides multiple opportunities for creative thinking. What play are they doing? What does shape-shifting look like? And, how does Papa Frita steal the spotlight?

In addition to filling in vague plot lines, students also started to use pantomime and/or create their own props from whatever they could find in the room. One group turned a sweatshirt into a treasure chest and hid markers (i.e., gems) inside the folds. Another group created a bridge out of string on the ground. And, of course, pencils were wands, and rhythm sticks were swords. Throughout this lesson, students demonstrated their divergent and convergent thinking abilities, yet they often did so when there was a concrete starting point. This provides evidence for the Green Eggs and Ham Hypothesis (Haught-Tromp, 2017), which is named after Dr Seuss’s challenge to write a book using only 50 unique words. Other existing work found constraints can be helpful for creativity (see Tromp, 2023 for an overview). By providing the goal (e.g., “navigating obstacles”), students can elaborate and build upon those words to create their world, but without that prompt, students may have been overwhelmed with all the possibilities. In this regard, Chat GPT3.5 allowed students a constrained space for their creativity to flourish.

With the script being developed externally, the student directors had more time to plan out their scenes, and when the rehearsal process began, student directors did not require much spontaneous in-the-moment planning. This freedom enabled them to dedicate their full attention to maintaining the actors’ focus, breathing life into their envisioned production, and skillfully orchestrating the entire rehearsal process.

Our students may not have been in creative flow in the beginning of this lesson, but when they started dissecting their scripts, conceptualizing how their stories would be told, their imaginations were both free and engaged. These students were incredible, thinking of ideas we would not have found in existing scripts, even if we looked for another year.

Limitations

Despite demonstrating proof of concept, this work contains several limitations, including a small sample size. Further, some study conditions may make it challenging to generalize. For example, having 3 instructors at each session is unlikely to happen in most educational spaces. However, we were also facing some unique challenges associated with teaching afterschool, such as waning attention spans and lack of authority status.

Specialized Support

An extra challenge was translating the complex components of students’ scripts to the stage, like battle scenes and shape-shifting powers. Our team was fortunate because one of our instructors was a trained choreographer, and he provided detailed guidance on how to choreograph multi-step fight sequences. While narrative texts allow readers to imagine fight scenes, playscripts require depicting the action physically for the audience (without hurting anyone in reality). The instructors should intentionally learn how choreographers develop these scenes.

Integrating Additional Assessments

One additional limitation is that we relied on previous observational/student journal data for our pre-assessments and their journals and final products as their post-assessments. Thus, we can report that the students developed a greater ability to discuss characters, settings, and plot points, but in future studies, we plan on adding additional measures, like standardized assessments.

Ethical Use of AI Language Models

Finally, we did not discuss the ethical questions surrounding ChatGPT with the students. We took the students’ prompts and returned with scripts with “help from computers,” but we did not directly review the ethics behind AI use. As an afterschool program, we did not think we had the time or authority to provide this level of guidance. If we completed this lesson in our own classrooms, however, we would discuss several key components of ethical AI use, establishing transparency with (a) families, (b) teacher expectations, and (c) ethical use. First, before using AI with students, we would need to discuss our policies and plans with the students’ families to review privacy/data security concerns and classroom policies for age-appropriate use. These policies would differ based on students’ developmental levels.

Next, we would be transparent regarding our expectations of when students can use AI and when it is inappropriate. For example, the goal of this lesson is not to teach students how to write a play script. If that were the case, we would not allow AI use. Teachers need to be very clear on the goals and objectives before deciding to use (or not use) AI support. Finally, we would establish a transparent policy for ethical use. Many academic journals are developing clear policies on how to interface and publish with AI language models. Although some journals are simply banning AI model use, most others have developed a more moderate approach. For example, the Journal for the American Medical Association (JAMA, 2024) maintains the following policy:

If these models or tools are used to create content or assist with writing or manuscript preparation, authors must take responsibility for the integrity of the content generated by these tools. Authors should report the use of artificial intelligence, language models, machine learning, or similar technologies to create content or assist with writing or editing of manuscripts in the acknowledgement section...

This policy could form a classroom contract with AI models. If students use AI models, they must acknowledge it within the paper and take responsibility for the accuracy/integrity of the content. Teachers would also need to follow this policy.

These ethical issues must be addressed with students because learning how to ethically use AI models to enhance and support their work is essential. Companies are now expecting new employees to enter the workforce knowing how to use these tools to do their jobs more efficiently, and it is important for our educational system to prepare students for that world (Watson & Stachowiak, 2024). If we can leverage this technology to also differentiate and engage students, then everyone benefits.

Future Work

In the future, we plan on addressing the limitations above and making several other adaptations. Given many students struggled to brainstorm the initial settings, characters, and plot points, we plan to adjust some of those prompts. Several studies have demonstrated different prompts can increase creativity on divergent thinking tests, so instead of asking for one setting, we could ask for one setting “no one else would think of” or simply, “be as creative as possible” (Said-Metwaly et al., 2020) We could also provide prompting questions, like what are five mysterious places you have seen or what is the most unusual place you would find a treasure.

Beyond empowering students, ChatGPT has empowered us, as instructors, and studying/replicating these experiences could inspire an additional line of work. As the objectives for the other modules (acting and designing) were different, we were planning on identifying one script for all students to anchor the development of acting and designing process skills. Yet, we still wanted to infuse high interest, appropriately leveled, and relatively short, and personally relevant scenes to use throughout the Brilliant Storyteller modules.

One novel we considered was Sideways Stories from Wayside School (Sachar, 1988) because each chapter was relatively short and silly. Different groups could do different chapters, and yet the whole group produced a coherent vision of the school. We couldn’t find the perfect, short, engaging script from this novel, so we copied a chapter into ChatGPT and asked for a script that met our criteria: 3 minutes, third grade reading level, with only three characters, no narrator, and more silliness. We had a good script in seconds, and then, we edited it further. For example, we changed all characters’ names to gender neutral so any student could play any character, we added additional funny ways to frustrate the teacher, and we incorporated a lot more hilarious (or groan-worthy) puns. We also adjusted the scene divisions ChatGPT created, condensed the action into longer chunks of text, and removed excessive repetition.

We replicated this with different versions of Cinderella, from the Chinese version, Yeh-Shen (Louie, 1996), to the Caribbean version, Cendrillon: A Caribbean Cinderella, (San Souci & Pinkney, 1998). We even developed a script from a Big Foot version (Thaler et al., 1997). Any story could serve as the anchor for a reader’s theater, which will expose students to endless culturally relevant, engaging scripts. Collectively, we, as instructors, are not playwrights and turning these stories into scripts would have taken us a significant amount of time, and in this way, ChatGPT empowered us as well.

Broader Implications and Conclusion

Beyond the theater application, it is important to recognize how AI can be leveraged to support all students in the classroom in unique ways. AI language model’s (e.g., ChatGPT) key advantage is its ability to dynamically generate customized content while adhering to specific parameters, enabling true curriculum differentiation. This can be used for twice-exceptional students who may have a deficit in one area and strengths in others. For example, if they are passionate about dinosaurs and have dyslexia that prevents them from reading many in-depth texts, AI can translate complex text into more digestible content specific language, and other software options could also read it aloud to the student. This allows students to continue to challenge themselves with advanced content while struggling with reading. See Table 2 for multiple examples beyond reading/language arts.

Table 2.

Broader Implications for AI Use in the Gifted Classroom.

Goal	Sample Prompt	ChatGPT 4.0 Responses
Simplify complex topics or incorporate more advanced perspectives for gifted students	Please explain the special relativity to a third grader	Imagine you’re in a car that’s moving really fast down the highway. Now, let’s talk about special relativity
		• Speed of light. Think of a super-duper fast light beam that travels faster than anything else. No matter how fast you’re moving in your car, if you turned on a flashlight, the light from it would always zoom away at the same super speed
		• Time slows down. If you had a magical clock and you were moving super-fast in your car, your clock would tick slower compared to someone standing still. It’s like time is taking a nap because you’re moving so fast!
Add lesson content, examples, and/or analogies based on students’ interests and background	What are three examples of things students be unable to do if they did not understand fractions?^a	Measurement: Measuring lengths, weights, and volumes. Whether you’re using a ruler, measuring liquids, or weighing ingredients, fractions are essential
		Interpreting data: Understanding charts, graphs, and statistics often involves fractions. For instance, reading a pie chart or understanding survey results requires knowing how to interpret parts of a whole
		Financial literacy: Managing finances, such as understanding interest rates, calculating mortgage payments, or dividing assets, often involves fractions and percentages
Develop problem-based learning experiences aligned with course content	Develop five problem-based learning experiences that help students to understand how historical events have shaped the modern world^b	Here are 2/5 examples for space
		Environmental impact of the industrial revolution
		Problem: How can we mitigate the environmental damage caused by industrial practices that began during the industrial revolution?
		Preserving cultural heritage
		Problem: How can we preserve the cultural heritage of communities affected by historical events like colonization and migration?

^aNow, as a teacher, we can select make one of these the anchor of a fraction lesson that allows students to discover why fractions are important. Allow students to try some of these scenarios without the option of fractions and discuss the importance.

^bFor this prompt, we needed to work a bit more with ChatGPT to clarify. We needed to add: “Problem-based learning requires an authentic problem.”

Overall, this action research study highlights the promising potential of using AI language models like ChatGPT to empower students as creators and facilitate differentiated, engaging literacy experiences. By leveraging ChatGPT’s capabilities, we efficiently generated tailored scripts that sparked student enthusiasm, provided authentic reading practice opportunities, and opened creative avenues for students to translate their imaginative worlds into performable stories. We gave students the tools to transform the worlds they have spinning in their minds into something real, something they can see, touch, and perform.

Despite some practical limitations, this approach worked. Students were reading more deeply and thinking more creatively. They were engaged, even afterschool. Moreover, it serves as a model for how educators can utilize AI to craft culturally relevant, student-centered learning materials customized to individual needs and interests. As AI tools continue advancing, further research is warranted to explore refining prompting techniques, addressing ethical considerations, and expanding applications across the curriculum. Ultimately, thoughtfully, integrating AI could transform students from passive consumers to active creators, bringing about a revolution in education, one that empowers every student to be the innovator, their own storyteller.

Footnotes

Acknowledgments

We would like to acknowledge the incredible team of people that made this project happen. Angela Wheeldon is our administrative coordinator, who provided financial organization. Ari Glover, Ruby Heller, and Anna Jones are the undergraduate, performing arts education majors, who supported the third and fourth graders through these adventures. Kathrin Maki, a reading interventionist researcher, helped integrate reading strategies into the curriculum. Finally (yet absolutely essential) were our YMCA community partners, including: Elsbeth Fritz, Hannah Caccamise, and Teagan Hayes. We are grateful!

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Jacob K. Javits Gifted and Talented Students Education Program of the Office of Elementary & Secondary Education at the United States Department of Education under award number [PR/Award # S206A220014].

ORCID iD

Lisa DaVia Rubenstein

Bios

Lisa D. Rubenstein, PhD, serves as a professor in the Department of Educational Psychology and the Director of the Center for Gifted Studies and Talent Development at Ball State University. Lisa is the principal investigator for Project Brilliance, which is designed to identify and develop creativity and leadership talent in students that may otherwise be overlooked.

Andy Waldron, PhD, is an Assistant Professor of Theatre Education at Ball State University. His current work focuses on theater education training and pedagogy, Theatre for Young Audiences, queer youth theater, and socially engaged arts practices. Recently, as lead researcher, he launched the LGBTQ+ Theatre for Youth Project; a Web site documenting, analyzing, and cataloguing queer youth theater.

Gerardo Ramirez, PhD, serves as an associate professor in the Department of Educational Psychology at Ball State University. Gerardo’s research focuses on examining the impact of frustration, empathy, and cultural capital on students’ success and persistence.

References

Besnoy

K. D.

(2023). ChatGPT’s ethical delima. In Teaching for high potential (pp. 5–8). Available at: https://nagc.org/global_engine/download.aspx?fileid=FCDE5687-B6A2-4C10-B99F-F53337E633D6&ext=pdf

Cohen

(2002). Acting one (4th ed.). McGraw Hill.

Firmender

J. M.

Reis

S. M.

Sweeny

S. M.

(2013). Reading comprehension and fluency levels ranges across diverse classrooms: The need for differentiated reading instruction and content. Gifted Child Quarterly, 57(1), 3–14. https://doi.org/10.1177/0016986212460084

Haught-Tromp

(2017). The Green Eggs and Ham hypothesis: How constraints facilitate creativity. Psychology of Aesthetics, Creativity, and the Arts, 11(1), 10–17. https://doi.org/10.1037/aca0000061

Journal for the American Medical Association (JAMA) . (2024). Instructions for authors. JAMA. Available at: https://jamanetwork.com/journals/jama/pages/instructions-for-authors#SecAuthorshipandDisclosures

Louie

A. L.

(1996). Yeh-shen: A Cinderella story from China. Penguin.

Morrison

(1981). Speech to the Ohio arts council. As quoted in E. Brown (1981, September 27). Writing is a third career for Morrison. The Cincinnati Enquirer, p. 106. Available at: https://apnews.com/article/7631ae6223894408b4fca49ab1874f4f

Morrison

(1992). Jazz. Knopf.

National Association for Gifted Children . (2024). Definition of asynchrony. National Association for Gifted Children. Available at: https://nagc.org/page/glossary

10.

Sachar

(1988). Sideways stories from Wayside wchool. Harper Collins.

11.

Said-Metwaly

Fernández-Castilla

Kyndt

Van den Noortgate

(2020). Testing conditions and creative performance: Meta-analyses of the impact of time limits and instructions. Psychology of Aesthetics, Creativity, and the Arts, 14(1), 15–38. https://doi.org/10.1037/aca0000244

12.

San Souci

R. D.

Pinkney

(1998). Cendrillon: A caribbean Cinderella. Simon & Schuster.

13.

Thaler

Elsie

Smeldred

K. Q. M. P.

(1997). Cinderella bigfoot. Scholastic.

14.

Tolan

S. S.

(1994). Giftedness as asynchronous development. Tip Network News, 4(1), 1–7. https://www.stephanietolan.com/gt_as_asynch.htm

15.

Tromp

(2023). Integrated constraints in creativity: Foundations for a unifying model. Review of General Psychology, 27(1), 41–61. https://doi.org/10.1177/10892680211060027

16.

VanTassel-Baska

Hubbard

G. F.

Robbins

J. I.

(2021). Differentiation of instruction for gifted learners: Collated evaluative studies of teacher classroom practices. In Smith

S. R.

(Ed.), Handbook of giftedness and talent development in the Asia-Pacific (pp. 945–979). Springer.

17.

Watson

C. E.

Stachowiak

(2024). Thinking with and about AI (No. 517) [Audio podcast episode]. In Teaching in higher education. Innovative learning. LLC. Available at: https://teachinginhighered.com/podcast/thinking-with-and-about-ai/

18.

Weston

(2021). Directing actors: Creating memorable performances for film and television. M. Wiese Productions.

19.

Ziernwald

Hillmayr

Holzberger

(2022). Promoting high-achieving students through differentiated instruction in mixed-ability classrooms—a systematic review. Journal of Advanced Academics, 33(4), 540–573. https://doi.org/10.1177/1932202X221112931