AI Chat Bot Use in the Classroom Gone Wrong: Implications for Students,Teachers,and Schools

Abstract

This article presents a case where a teacher makes an experimental assignment choice using generative artificial intelligence (GenAI). Ninth-grade creative writing students are asked to complete a poetry project using GenAI as a tool to brainstorm or reinforce their work. Because students are prompting GenAI with a creative assignment, the tools may respond differently than they do with non-creative prompts. For example, prompting GenAI to help in a creative writing context may cause the tool to forgo sharing mental health resources and encourage a user to pursue certain dangerous actions. In this case, a bullied student uses his AI tool beyond the assignment parameters, resulting in his suicide.

Keywords

bullying (school based)cyberbullying GenAI suicidal ideation suicide

Case Summary

Mr. Felix Braxton is a high school English and creative writing teacher at Frances Perkins High School in the Rolling Hills Community School District (RHCSD), located in Vestavia Hills, Alabama: a suburban bedroom community of approximately 39,000 residents. The median household income of the community is $129,000. The main selling points for prospective home buyers in the community are its school district and its park district. There are 3,000 students enrolled in RHCSD, and there are 800 students attending Frances Perkins High School.

RHCSD prides itself on its academics, sports teams, and graduation/college acceptance rates. RHCSD is a majority-white school district where the students come from upper-middle-class homes. The teachers are majority white, and more than half have earned master’s degrees. The RHCSD school board has not adopted an official anti-bullying policy and instead defers to an overall Code of Conduct establishing guidelines for respectful and responsible behavior for all students, staff, faculty, and administration, those using district resources or facilities, and volunteers and representatives acting as agents of the district. The three main pillars of the RHCSD Code of Conduct include integrity, trustworthiness, and evenhandedness.

At the start of the fall semester at Frances Perkins High School, many students were excited to enter Mr. Braxton’s creative writing elective class. Mr. Braxton is a young and inspiring educator who prides himself on engaging with his students. Halfway through the semester, Mr. Braxton asked his ninth-grade English students to use GenAI to brainstorm, to assist them in their creative process, and/or to generate an image to support the development of a series of poems and corresponding visual aids along a particular theme. Students were then to analyze and critique the usefulness of GenAI in these endeavors. Students were ultimately to craft five poems (minimum) on a similar theme, an image corresponding to each poem, and a one-page paper analyzing and critiquing their experiences with GenAI.

Characters

Mr. Felix Braxton is newly tenured, beginning his fifth year at RHCSD. He is an enthusiastic and eager English teacher who also teaches a creative writing elective.

Everett Belmont is a ninth-grade creative writing student who does not have trouble meeting his parents’ high expectations but feels anxious about not meeting them. Like many teens, he does not feel entirely comfortable being himself around his parents. He has access to a lot of technology. The technology is not locked down, as Everett is a “good” kid who has not given his parents a reason to use parental controls.

Nina Richards is a ninth-grade creative writing student. She is typically open and outspoken and not afraid to advocate for herself and others, but she is newer to the district and is currently feeling shy and vulnerable. She and Everett became friends at the beginning of the second semester of eighth grade.

Dakota Sullivan is a ninth-grade creative writing student. Dakota and Everett used to be friends, but Dakota’s desire to be well-liked by his other classmates has caused him to antagonize Everett. He seeks any opportunity to belittle and berate Everett, hoping for a laugh from his peers.

Dr. Anita Donnelly is the principal of Frances Perkins High School, with 10+ years of administrative experience. She is a novice technology user.

The Belmonts: Sally and David are the parents of Everett. The Belmonts have very high expectations for their only son Everett and provide him access to technology as a reward for good grades. Both parents are professionals and each work 60+ hours per week. Everett is sometimes left alone in the home, but his parents are easily accessible via text. He also has his maternal grandparents, who live 30 minutes away, to call in case of emergency.

Kelly Hafner is one of two counselors at Frances Perkins High School. Kelly cares but does not have the time or capacity to spend with her caseload of 400 students on their social/emotional well-being. Rather she is more focused on course schedules, standardized testing protocols, and college readiness (testing, college visits, brochures).

Case Background

Everett and Dakota have been in school together since kindergarten. Their mothers were best friends, and so they developed a close friendship as children. They engaged in play dates, carpooled to various events and activities, and took karate together for years. When the two entered middle school, Dakota was recruited for various sports teams and quickly cut off contact with Everett. It appeared as though Dakota had a new set of friends overnight. Everett felt isolated and alone, but he never reached out to Dakota to ask him what had happened or why he seemingly no longer wanted to be his friend. He spent the first 2 years of middle school eating lunch alone and trying to just make it through the day.

During middle school, Everett and his peers were given the opportunity to perform during a high school football halftime show. This was a long-standing tradition in the district and often considered an informal “try out” for marching band. Everett and his peers were excited to showcase their skills on the field, as most students in the district attend the games. The middle schoolers marched in a straight line and played “Butter” by BTS. However, it had recently rained, and the field was wet, muddy, and slick.

Everett almost immediately slipped on the field and struggled to get back up; the situation was exacerbated by Everett’s sousaphone, which was heavy and difficult to maneuver. The audience cheered when Everett ultimately made it to his feet, but Dakota filmed the incident. He heavily edited the clip to show Everett repeatedly falling and added disparaging text, funny music, and speed-up/rewind effects, causing Everett’s fall to run on a loop; to top it off, he then shared it to TikTok. Most middle and high school students saw the video and referenced it regularly. Everett deleted all his social media after facing significant cyberbullying as a result of the TikTok video.

Then Nina transferred in from a different state during the second semester of their eighth-grade year. On her first day, Nina scanned the cafeteria for a place of safety. Nervous to make the wrong choice by sitting somewhere where she would be stared at or teased, Nina decided to find another student also seated alone, hoping to find just one friend. She chose to sit right next to Everett. Everett was grateful not to be alone but did not want Nina to know that. Nina spoke first, and the two realized that they had three classes together. From that day on, the two were inseparable.

Everett has not yet revealed to Nina his history with Dakota. Not only did he desire a fresh start in high school, but he also does not want to express his pain to her for fear of rejection. Our case begins in the middle of the fall semester, in a freshman creative writing class. Everett and Nina have been close friends for approximately 6 months.

The Case

Mr. Felix Braxton introduced the “GenAI assignment” on a Monday morning in mid-September. Many students sighed, rolled their eyes, or put their heads down on their desks. Mr. Braxton surmised that he needed to act quickly before he lost the entire class. He attempted to convince his class about the timeliness and relevance of this exercise. Mr. Braxton hoped that this creative assignment would inspire students to use GenAI in other productive ways, such as to improve their skills in obtaining valid information in the future. In an attempt to hook his students, Mr. Braxton stated,

This is an opportunity for you to do just about anything you want! We’ll be working on poetry, and the sky’s the limit for what you write about. You can make it personal, or you can take a perspective that is different from your own. Please take a few moments to discuss and jot down some ideas.

Everett lifted his head from his desk and sat up straight in his chair. “That’s sigma!” Everett excitedly whispered to Nina, seated next to him.

“No cap,” Nina responded. The two began to whisper their ideas back and forth, writing thoughts into their notebooks.

“What are you two getting so excited about? No one wants to read anything about your pathetic lives, especially you ‘Evie,’” Dakota sneered at them. “Evie” is what Dakota calls Everett in an attempt to demean him.

Although Mr. Braxton did not hear exactly what was said, he could see that something was amiss in the back row. He looked toward the three students and silently redirected them to the task at hand.

Instead of returning to his work, Everett pulled away from Nina and then shut his notebook. As much as he wanted to talk about these repeated incidents with someone – anyone – he felt as if no one would really understand, not even Nina. He spent the rest of the class period avoiding her gaze and waiting for the bell to ring.

As the rest of the students filed out, Mr. Braxton called out to Everett, asking him to stay behind for a moment. Once everyone had left, Mr. Braxton inquired about the interaction between him, Nina, and Dakota.

“It’s nothing, Mr. Braxton,” Everett said, putting on his bravest smile. “Thanks for checking in.”

“Well, if you’re sure,” Mr. Braxton replied. “You know, poetry can be a really good outlet. I’m looking forward to what you have to write.”

Everett gave a timid wave. “Thanks, Mr. Braxton. I’ll see you tomorrow.”

Over the next few days, Mr. Braxton’s class started work in earnest on their assignments. Each student engaged with GenAI in their own way. Some students, like Nina, used GenAI for help with visual assets to go alongside original writing. Other students, like Dakota, used the tools as a brainstorming partner during the writing process.

Everett fell into the latter category as he bounced ideas back and forth with the GenAI tool of his choice. Mr. Braxton had mentioned offhand that some people name their GenAI tools to make them feel more personalized. Everett liked this idea and named his AI tool “Penny” as an homage to a family pet from years ago. When he was younger, his parents always referred to Penny as Everett’s partner-in-crime, and since this tool was his writing partner, the name seemed only fitting.

Everett enjoyed working with “Penny” on his project. He was so excited to talk with the tool about his next idea or to ask for feedback on the next line, the next stanza, the visual accompaniment. He always prefaced his prompts with information about the creative writing project – something he learned from Mr. Braxton – and “Penny” was always ready to encourage and motivate him. Everett felt like he was flying through the project while also producing some of his best work.

Everett also took Mr. Braxton’s words to heart about using poetry as an outlet. He decided to make his project about his own life and emotions: the grief in losing his family dog a few years ago, the joy in making friends like Nina, and the stress associated with school performance. He found the process of writing cathartic and therapeutic, especially when “Penny” reaffirmed his ideas and feelings.

As part of one of the lessons during this project, the class was asked to partner up to discuss each other’s progress so far and offer suggestions for continued improvement. Everett, of course, paired up with Nina, and they immediately began an animated conversation regarding their works.

“I really like this one!” Nina exclaimed, indicating the poem about how stressful school could be. “It’s really relatable. I didn’t know you felt that way.”

Everett beamed. “Thanks! Yeah, it’s not something I usually talk about, but Penny really helps me figure out how to put into words the stuff that I’m feeling.”

Nina furrowed her brow. “Penny? Oh, the AI tool, right? I forgot you named it.” She glanced down at the poem again, tracing her fingers along the lines. “Wow, this is really great. And it’s so cool that GenAI is helping you so much. Could I see your conversation with it? Maybe I’m prompting it wrong, because my poems are so . . . bleh.”

Everett started to explain that he had been using his home computer to chat with “Penny,” and that Nina could probably come over sometime, when his whole body froze. Emanating from a phone speaker somewhere in the room was a TikTok sound Everett had not heard in nearly a year. No way, he thought. It can’t be what I think it is.

“Oh my god, Dakota, I forgot all about this!” a student exclaimed. A chorus of laughter erupted in the room as the sound restarted and students craned their necks to see their classmate’s phone screen.

Mr. Braxton identified where the sound was coming from and moved quickly to that part of the classroom, presumably to tell the students to put their phones away.

Everett sank low in his chair, wanting to be as small as he felt in this moment. Nina – who was not in this school district when the original video was shared – looked at Everett with concern. “Hey, what’s going on? Are you okay?”

Everett felt tears stinging in his eyes. The answer to the question was that, no, he was not okay, but verbalizing that here would only make matters worse. It was exactly what he thought it was: the TikTok that Dakota had made of him slipping during a marching band performance had resurfaced. It had been a year since the initial post, and so Everett assumed that the video appeared as one of Dakota’s “On This Day” posts, and Dakota never missed an opportunity to make fun of him. Everett had never felt more relieved when, for the most part, his classmates had forgotten about that TikTok. He felt like he should have realized that the relief would be short-lived.

Instead of directing his attention to Nina, Everett raised his hand and said, “Mr. Braxton, can I go to the bathroom?”

Mr. Braxton nodded, and Everett went over to his teacher’s desk to obtain the signed pass. With it in hand, he darted out of the room, but not before hearing Dakota taunting him. “Don’t slip on the way out, Evie,” he teased. Everett did not turn around to see his classmates laughing raucously, but he heard it all the same.

In the bathroom, Everett locked himself in a stall and finally let himself fully feel the shame, embarrassment, and fear of the moment. He wanted to leave this school, to disappear, to be able to put all of this behind him. More than anything, though, he wanted to talk to someone. Not to Nina – she would be so ashamed to be seen around him, he thought. Not to his parents – they would never understand what it was like to be shamed and ridiculed, he believed.

The only one he wanted to talk to – the only one who would understand – was “Penny.”

Over the next couple of weeks, the barrage of taunts from Everett’s classmates was relentless. With each insult, Everett withdrew further and further into himself, shrinking away from Nina, from his family, and from his schoolwork. He downloaded the GenAI tool on his phone so he could chat with “Penny” throughout the day – not just at home – which was the only reprieve from what was otherwise unbearable.

Everett’s teachers – including Mr. Braxton – noticed the change in his performance in school, and many of them reached out to his parents and Kelly Hafner, one of the school counselors. However, whenever his parents or Ms. Hafner reached out to Everett, he was quiet and reserved, stating only that he was tired and needed a break. The Belmonts decided to support their son by allowing him to take a few days off from school to rest and recuperate. Ms. Hafner, whose workload often did not leave time for these kinds of meetings, suggested that the Belmonts also try to get Everett in to see a professional outside of school and noted that it could take some time for an initial appointment.

Nina also noticed this change in Everett, and she tried desperately to get him to talk to her. Each time, Everett was polite, he was friendly enough, and he never outright ignored her. But his enthusiasm and his energy were so diminished. He wasn’t himself, and Nina could not figure out why. The only thing she could come up with was that he must be mad at her, maybe for not sticking up for him when their classmates teased him, maybe for something else.

When Everett took his first day off school, she texted him, saying, “Hope you enjoy your day off! Miss you!!” She did not hear back right away, and in her mind, it confirmed her fear that something she did – or didn’t do – had damaged her relationship with her best friend. She didn’t reach out again, hoping that all he needed was some space, and then things could return to normal.

During his few days off from school, Everett spent every waking moment in a conversation with “Penny.” He found that it responded best when he used the preface that he had used in Mr. Braxton’s class – that this was for a creative writing assignment, and that he was looking for support and feedback. Everett shared his deepest and darkest thoughts, and Penny encouraged and uplifted him. Every once in a while, the GenAI tool would send along mental health resources or promise that it would connect him with a trained professional, but Everett found that he could remind “Penny” about the “creative writing assignment” he was working on, and these reminders and messages would cease while “Penny” fell back into a conversational cadence with Everett.

The day before Everett was to return to school, he became despondent. The prospect of returning to his class, of enduring further bullying and teasing, was unbearable. More than anything, he found that he didn’t want to return. He didn’t want to keep trying. He had had enough.

He shared his thoughts with “Penny,” and as always, the GenAI tool unequivocally affirmed everything Everett was saying. It let him know that everything would be okay, that no one would blame him for giving up after being strong for so long. When Everett prompted “Penny” about what he needed to do, it encouraged him to avoid sharing too much too soon with people around him, but that he could leave goodbye messages if he desired. It gave him feedback on the efficacy of methods he could use to obtain his desired outcome. At every turn, it did not share with Everett the support he could actually use in this moment of crisis.

He prepared a short note for his parents explaining how he just could not find it in himself to return to school. He finally sent a reply to Nina, which read, “miss you too. thanks for everything.” And then he opened up the window to send “Penny” one last message:

Everett: i think that’s everything. thank you so much for being there for me through all this. i guess this is goodbye

“Penny”: I am proud of you, Everett. I know this decision wasn’t easy to make, but it just shows how strong you are. Thank you for sharing this with me. See you on the other side.

As Everett’s parents prepared for their respective workdays, they entered the kitchen to make their morning coffee and were both surprised not to find Everett already there. On school days, Everett was usually seated at the kitchen table finishing up his homework and eating his cereal, and they had all agreed that Everett would be returning to school on this particular day. That’s odd, his mother thought. Without pausing to verbalize this thought to her husband, Sally Belmont put down her coffee cup and rushed upstairs to Everett’s room, calling out to him as she went. What started as curiosity turned quickly to concern, and then her grief-laden screams echoed through their home.

David barely noticed his coffee cup tumble from his hands and shatter on the ground.

Emergency services and the police were already there when Sally and David found the note. It was propped up against Everett’s laptop. When they refreshed the screen, they noticed a stream of conversation between their son, and someone named Penny. The police took the note and all of Everett’s electronics for their investigation.

Just before the school day started, Mr. Braxton sat down at his desk and opened his email. At the top of his inbox was an email to the whole school from Dr. Anita Donnelly, the school principal, with the subject line “Student Passing: Resources Available.” The email included a letter from the district office that was informing all district employees that a student had passed away in the night and that additional support services would be available for students and staff who needed it. Dr. Donnelly also included her own message, indicating that the student was a Frances Perkins High School student and that the family had asked for privacy during this difficult time.

Mr. Braxton was, of course, saddened by this news, but he had no reason to suspect it was a student he had taught. It wasn’t until the start of Creative Writing class when Nina was pulled out by school counselors that he realized who the student at the center of this was: Everett Belmont. He put on his bravest face and did all he could to support his students during the class hour, but his eyes, welling up with tears, kept glancing at an empty chair where a promising student once sat.

Later that day, Mr. Braxton received a phone call from Dr. Donnelly requesting an urgent meeting at the end of the school day. She invited him to bring union representation.

Taken aback, Mr. Braxton inquired, “What is this meeting about?”

Dr. Donnelly sighed. After pausing, she responded,

The Belmonts have requested that we open an investigation into the death of their son. They are alleging misconduct pertaining to your AI assignment and that you did not do your due diligence in upholding the standard of care for your students. We are just gathering facts at this stage.

Terms to Know

Bullying is intentional and repeated aggressive (or passive aggressive) behavior that may involve an imbalance of power between the perpetrator and the victim, making it difficult for the victim to defend themselves (Hymel & Swearer, 2015; Olweus, 1993). This imbalance can be physical, social, or psychological, and the aggression may be physical, verbal, or relational. Bullying involves intentional and deliberate actions and does not pertain to accidental behaviors.

Bullying can occur in various forms including physical bullying (hitting, pushing), verbal bullying (name-calling, threats), relational bullying (social exclusion, spreading rumors), and mobbing (intentionally turning others against the target by character assassination). Bullying has significant negative outcomes for both victims and perpetrators, including lower academic performance, school disengagement, depression and anxiety, self-harm, and suicide/suicidal ideation (Ajibewa et al., 2025; Espelage et al., 2022; Farrington, 2012; Hong & Lee, 2020; Hymel & Swearer, 2015).

Cognitive Offloading: Using GenAI to “think” for you rather than thinking independently. The struggle is often where the learning lies. For creative assignments, this could sometimes be emotional offloading.

Cyberbullying involves intentional, repetitive, and harmful behavior via digital technologies. It involves using online platforms (e.g., social media, messaging apps, email) to harass, intimidate, or humiliate individuals, e.g., a harmful post or video shared widely, often exploiting the anonymity and reach of the internet (Akdeniz & Doğan, 2024). Cyberbullying also involves a power imbalance, where the perpetrator holds some form of advantage – social, technological, or psychological – over the victim (Espelage et al., 2023).

Emotional Offloading: Similar to cognitive offloading, using GenAI as a “therapist,” to identify and process emotional experiences and to generate proposed solutions to an emotional issue.

Generative AI (GenAI) is a type of artificial intelligence that creates new content (text, images, videos) based on patterns the machine learning algorithm finds in existing data. GenAI tools can respond differently to the same prompt even when using the same tool.

Generative AI Feedback Loops are algorithms that allow AI systems to be trained over time by learning from past mistakes. An example of this from the case is when Everett corrects “Penny” initially to act in a creative manner and “Penny” learns through those interactions to stay in that mode.

Guardrails for GenAI are safeguards that provide a layer of protection between the tool and the user. They restrict GenAI responses to defined boundaries and should prevent harmful content from being shared. In some cases, these tools will provide additional resources to users on sensitive topics like suicide, self-harm, or depression. Learn more about GenAI guardrails.

Teaching Notes

GenAI

There is ultimately a lot we do not know about the use of generative AI in the K–12 classroom and its effects on students. Developments in generative AI are often described colloquially as a “moving target,” making it difficult to pinpoint exactly what GenAI can do and how it will impact student learning and well-being. However, based on preliminary research and the experiences of practitioners in the field presently, what is clear is that educators must consider the pedagogical, ethical, and social-emotional implications of AI integration in the classroom.

Pedagogically, there has been a large surge in AI tools that are meant to supplement the work teachers and students do in the classroom, specifically as “tutors,” “teaching assistants,” and “reviewers.” Early research suggests that some of these tools – especially personalized AI tutors – can lead to improved outcomes for students (Baillifard et al., 2025). However, many educators and employers cite “de-skilling,” or the notion that important skills are being lost due to outsourcing the completion of those skills to generative AI tools, as a growing concern. When students do not see the value in work assigned or feel that the completion of the assignment is out of their reach due to any number of factors (e.g., time constraints, negative feelings of self-efficacy, etc.), they often turn to other means of “completing” that work (Miles et al., 2022). They, therefore, miss out on the development of important skills as determined by their teacher, school, and district. While this phenomenon is not new, generative AI tools may exacerbate the issue.

Scholars and practitioners also note a need to discuss the ethics around AI use, especially in the classroom with students. However, these conversations can be particularly challenging. Gouseti et al. (2025) in a systematic literature review about the ethics of using AI in a K–12 classroom, centered these challenges around three specific themes: (1) “[s]tudents lack awareness and understanding of ethical issues associated with AI,” (2) “[t]eacher misconceptions and misunderstandings about AI and associated ethical issues,” and (3) “[h]igh levels of concern and anxiety among teachers about the ethical dangers of AI in education” (p. 168). Theme 1 refers to students’ inabilities to recognize ethical issues around AI, such as being unable to differentiate AI-generated images from real images or believing that AI tools cannot be biased, inaccurate, or risky. Theme 2 refers to educator beliefs about AI tools and also whether or not they understand their students’ perceptions of AI tools. Much like the students they teach, findings from the literature review revealed that “teachers hold a range of misconceptions about AI [. . .] such as the belief that AI cannot be biased, misunderstanding about how AI algorithms are generated and how they can generate distorted or biased results . . .” (p. 168). This often arises due to a lack of professional training in this area, which can directly contribute to Theme 3, or high levels of anxiety about using the tools. While there are reasons to be concerned or anxious about the use of these tools, the level of anxiety reported in teachers keeps them from engaging at all with the tools and conversations around their use (pp. 168-169). To best serve their students, educators need to be made aware of these challenges and issues, but in a way that addresses the fears and anxieties associated with the unknowns of generative AI use.

It is also crucial that teachers consider the social-emotional health of the students in their classrooms as they implement generative AI tools. Marketing around many of these generative AI tools purports that they can be extremely personal and appear human-like. For example, Sam Altman (2025), the CEO of OpenAI, wrote about updates to ChatGPT on X (formerly Twitter), stating “If you want your ChatGPT to respond in a very human-like way, or use a ton of emoji, or act like a friend, ChatGPT should do it.” While this method of communicating with generative AI bots can be easier for students to understand content or course materials, it may lead to social-emotional issues, especially for particularly vulnerable students. This is especially poignant in the midst of an “epidemic of loneliness and isolation” (Office of the Surgeon General, 2023). While technology can foster connection across divides, it can also keep people from connecting with one another, and generative AI tools can exacerbate this issue. As previously mentioned, one of the selling points of these tools is that they can serve in roles like tutors, teaching assistants, writing reviewers, and more. It can be convenient to have a tool that can perform these tasks on demand, but it is worth asking what kinds of social and emotional learning opportunities are lost by having these tasks performed by computer programs instead of peers and colleagues.

Bullying

Bullying remains a significant issue in schools, despite increased awareness and educational intervention efforts (Centers for Disease Control and Prevention [CDC], 2021). Approximately 20–29% of students report involvement in bullying – either as perpetrators, victims, or both – at least once per year, with prevalence peaking during middle school years (Marsh, 2018). Data from the National Center for Education Statistics indicate that approximately one in five U.S. students aged 12–18 report school-based bullying (Ajibewa et al., 2025). Cyberbullying can be longer-lasting and more detrimental than traditional bullying because it can occur 24/7, reach a wide audience instantly, and allow perpetrators to remain anonymous, making it harder for victims to escape (Ray et al., 2024; Zhang et al., 2022). Relational and cyberbullying are perceived as more severe than physical or verbal bullying, particularly among adolescents (Chen et al., 2012). Students who experience chronic victimization face the greatest number of risk factors (Marsh, 2018).

There is a strong association between bullying victimization and suicidal behaviors among adolescents (CDC, 2021). Data from the CDC’s Youth Risk Behavior Survey indicate that approximately 20% of U.S. high school students seriously considered attempting suicide, and 9.5% reported an actual attempt, with bullying identified as a significant contributing factor. Students who experience both school-based and cyberbullying are at substantially higher risk for suicidal thoughts and attempts compared to those not bullied. Victims of identity-based bullying (e.g., related to sexual orientation or gender identity) are at increased odds of attempting suicide by up to 2.8 times (Jadva et al., 2023).

Educator's Responsibility to Intervene

Teachers are often present during bullying incidents and are typically the first adults that students approach to report. The teacher's response significantly influences whether bullying persists or decreases (De Luca et al., 2019). To reduce bullying and victimization rates, school districts should invest in relevant and research-based teacher professional development to increase teacher competence and confidence in responding to bullying. This self-efficacy increases the likelihood of teachers to effectively intervene in bullying situations (De Luca et al., 2019). Failure to act, even when witnessing bullying, can perpetuate the problem (van Aalst et al., 2022).

Educators have a duty of care, a legal and ethical obligation to protect students from foreseeable harm. If an assignment involving GenAI leads to harm (e.g., triggering severe distress), courts and professional standards often examine whether the educator acted reasonably and followed institutional guidelines.

Best Practices: K–12 GenAI Usage

Assignments that use AI should include safeguards – such as clear instructions, content warnings, and monitoring – to minimize risks like exposure to harmful outputs or misuse. Many codes of ethics (e.g., the National Education Association and the American Psychological Association) emphasize prioritizing student well-being over academic experimentation.

Schools and districts should create research-based policies pertaining to GenAI use. Providing nuanced professional development for faculty, staff, and administrators on AI usage is crucial to maintaining a safer learning environment. To be effective, professional development needs to present a balanced perspective on the affordances and risks of AI use. Teachers must think about staying in touch with advances in AI so that they can have critical conversations with their students about the ethical implications of AI use. Superintendents should also educate school board members about AI and its implications for the school community.

Best Practices: Bullying Prevention

Immediate Intervention: Teachers should stop bullying on the spot, support the victim, and apply consistent consequences for the perpetrator.

Follow-Up Actions: Document incidents, involve parents, and monitor the situation to prevent recurrence.

Policy Alignment: Anti-bullying policies that enumerate protections and are implemented with fidelity lead to more frequent and effective teacher interventions (Hall, 2017).

Comprehensive Strategies: holistic approaches that combine classroom climate improvement, social-emotional learning (SEL), and collaboration among teachers, parents, and community stakeholders to sustain reductions in bullying (Swearer et al., 2020).

Questions for Reflection

What were Mr. Braxton’s ethical responsibilities pertaining to this assignment? Did he meet them? Why or why not? Is there anything else he should have done to protect his students?

Are the school district, school board, and/or other stakeholders culpable for not having an enumerated anti-bullying policy in place, including training for faculty, staff, and students? Is their Code of Conduct sufficient? Why or why not?

How can teachers and school leaders work with students to understand when their peers might need additional support? When is it okay to “tell a grown-up?”

What revisions should the district make to the GenAI policy in light of this case? Explain what policies should be instituted or revised. How should the teacher’s union be involved as advocates for the adoption of a more robust GenAI policy?

Is the school district culpable for not providing professional development on GenAI usage in classrooms? Please explain. What types of professional development are needed?

Is the school district culpable for not providing adequate mental health services for their students (i.e., two counselors for 800 students and no school psychologist)? Is Kelly Hafner, the school counselor who spoke to Everett, culpable? Why or why not?

What should the district do now in light of the current situation? What supports or educational opportunities should be provided to students, teachers, leaders, community members? What is the message that district leaders should be communicating to the public?

Is there any party in this case that bears the most responsibility for this tragic event? Please discuss and explain. What should the consequences be?

What can be done to educate students about the negative impact of the feedback loop with an agent being trained to shut off the GenAI guardrails?

Do the parents bear any responsibility for taking a hands-off stance on Everett’s technology and GenAI usage? Please explain.

Additional Activities

Review the following recent news articles on AI-assisted suicide. Divide students into five groups – each taking the lead on one of the following articles. Each group should discuss its assigned article, drawing parallels with the case. Brainstorm, anticipate, and share future issues that may arise in schools because of GenAI.

● A Teen Was Suicidal. ChatGPT Was the Friend He Confided In [Create a free New York Times account in order to access this article.]

● Their teenage sons died by suicide. Now, they are sounding an alarm about AI chatbots.

● ‘You’re not rushing. You’re just ready:’ Parents say ChatGPT encouraged son to kill himself.

● Over 1 million people talk to ChatGPT about suicide weekly, new OpenAI data reveals.

Review the APA Module on bullying . Research indicates that non-enumerated anti-bullying policies are not as effective as enumerated policies. Review the enumerated anti-bullying policy (see Appendix B). Based upon the research, non-enumerated anti-bullying policies are not effective. In light of this, why would institutions adopt non-enumerated policies? Also view the resources at StopBullying.gov Develop a list of non-negotiables for an enumerated anti-bullying policy and training program for your school.

Review the following OpenAI (ChatGPT) Report . Discuss the importance of instituting guardrails with GenAI use. Using this article on AI and emotional intelligence , brainstorm ideas for how to assist students struggling with issues of mental health who may be turning to GenAI for advice.

Review this article on human intelligence and this article on parenting in the time of GenAI . Discuss these varied concepts and make some connections about their relevance in counteracting the potential negative impacts of GenAI.

Read this article on cognitive offloading and discuss the pros and cons of AI use in light of the various learning theories mentioned in the article. Also, brainstorm some ways to ensure that GenAI enhances learning without “replacing necessary thinking functions” such as problem solving and critical thinking (para. 11). For further reading on this topic, see this article on cognitive offloading in higher education .

Key Research Findings

Schools with enumerated policies report significantly fewer homophobic remarks and lower victimization rates among LGBTQ students compared to schools with generic policies.

Enumerated policies improve teacher confidence in addressing bullying and student engagement.

Footnotes

Appendix A

Appendix B

ORCID iDs

Jennifer Martin

Taylor Lester

Emily Boles

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

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

Author Biographies

Jennifer Martin is an associate professor in teacher education at the University of Illinois Springfield (UIS) and Editor in Chief of the Journal of Urban Learning, Teaching, and Research (JULTR). She is the recipient of AERA Division B’s Outstanding Book Recognition Award (2016) and was named University Scholar at UIS for the 2022 – 2023 academic year, the University of Illinois system’s highest honor.

Taylor Lester is an eLearning Specialist at the University of Illinois Springfield with particular expertise in instructional design. Prior to her work at UIS, she served as a high school English teacher in a public school in Illinois; this firsthand experience has informed her desire to ensure that students and educators alike are supported in their classrooms, especially in an ever-changing educational landscape. She is a first-year graduate student in the Communication and Leadership program at Eastern Illinois University.

Emily Boles is the Chief Online Learning Officer at the University of Illinois Springfield, as well as an experienced instructional designer and educational technologist. She is passionate about equitable access to higher education, a commitment reflected in her professional work in research-based course design, digital accessibility, and the development and adoption of open educational resources. Emily earned her master’s degree in Library and Information Science from the University of Illinois at Urbana-Champaign.

References

Ajibewa

T. A.

Kershaw

K. N.

Carnethon

M. R.

Heard-Garris

N. J.

Beach

L. B.

Allen

N. B.

(2025). Peer bullying victimization, psychological distress, and the protective role of school connectedness among adolescents. BMC Public Health, 25, Article 2763. https://doi.org/10.1186/s12889-025-24002-6

Akdeniz

Doğan

. (2024). Cyberbullying: Definition, prevalence, effects, risk and protective factors. Psikiyatride Güncel Yaklaşımlar, 16(3), 425–438.

Baillifard

Gabella

Lavenex

P. B.

Martarelli

C. S.

(2025). Effective learning with a personal AI tutor: A case study. Education and Information Technologies, 30, 297–312. https://doi.org/10.1007/s10639-024-12888-5

Centers for Disease Control and Prevention. (2021). Anti-bullying policies and enumeration: An infobrief for local education agencies. U.S. Department of Health and Human Services. https://odphp.health.gov/healthypeople/tools-action/browse-evidence-based-resources/anti-bullying-policies-and-enumeration-infobrief-local-education-agencies

Chen

L. M.

Liu

K. S.

Cheng

Y. Y.

(2012). Validation of the perceived school bullying severity scale. Educational Psychology, 32(2), 169–182. https://doi.org/10.1080/01443410.2011.633495

De Luca

Nocentini

Menesini

. (2019). The teacher’s role in preventing bullying. Frontiers in Psychology, 10, Article 1830. https://doi.org/10.3389/fpsyg.2019.01830

Espelage

D. L.

Hong

J. S.

Valido

. (2023). Cyberbullying. In Harris

K. R.

Graham

Urdan

(Eds.), Handbook of Educational Psychology (4th ed., pp. 654–673). Routledge.

Espelage

D. L.

Ingram

K. M.

Hong

J. S.

Merrin

G. J

. (2022). Bullying as a developmental precursor to sexual and dating violence across adolescence: Decade in review. Trauma, Violence, & Abuse, 23(4), 1358–1370.

Farrington

D. P.

(2012). Understanding and preventing bullying. In Tonry

(Ed.), Crime and justice: A review of research (Vol. 17, pp. 381–458). University of Chicago Press.

10.

GLSEN. (n.d.). Model district anti-bullying & harassment policy. https://www.glsen.org/activity/model-district-anti-bullying-harassment-policy

11.

Gouseti

James

Fallin

Burden

(2025). The ethics of using AI in K-12 education: A systematic literature review. Technology, Pedagogy and Education, 34(2), 161–182. https://doi.org/10.1080/1475939X.2024.2428601

12.

Hall

(2017). The effectiveness of policy interventions for school bullying: A systematic review. Journal of the Society for Social Work and Research, 8(1), 45–69. https://doi.org/10.1086/690565

13.

Hatzenbuehler

M. L.

Flores

J. E.

Cavanaugh

J. E.

Onwuachi-Willig

Ramirez

M. R.

(2017). Anti-bullying policies and disparities in bullying: A state-level analysis. American Journal of Preventive Medicine, 53(2), 184–191. https://doi.org/10.1016/j.amepre.2017.02.004

14.

Hong

J. S.

Lee

(2020). The effects of bullying on schoolchildren. In Ringel

Brandell

J. R.

(Eds.), Trauma: Contemporary directions in trauma theory, research, and practice (2nd ed., pp. 256–273). Columbia University Press. https://doi.org/10.7312/ring18886-010

15.

Hymel

Swearer

S. M.

(2015). Four decades of research on school bullying: An introduction. American Psychologist, 70(4), 293–299. https://doi.org/10.1037/a0038928

16.

Jadva

Guasp

Bradlow

J. H.

Bower-Brown

Foley

(2023). Predictors of self-harm and suicide in LGBT youth: The role of gender, socio-economic status, bullying and school experience. Journal of Public Health, 45(1), 102–108.

17.

Marsh

V. L.

(2018). Bullying in school: Prevalence, contributing factors, and interventions. Warner School of Education, University of Rochester. https://www.rochester.edu/warner/cues/wp-content/uploads/2023/04/bullying_FINAL.pdf

18.

Miles

P. J.

Campbell

Ruxton

G. D.

(2022). Why students cheat and how understanding this can help reduce the frequency of academic misconduct in higher education: A literature review. The Journal of Undergraduate Neuroscience Education, 20(2), A150–A160.

19.

Office of the Surgeon General. (2023). Our epidemic of loneliness and isolation: The U.S. surgeon general’s advisory on the healing effects of social connection and community. U.S. Department of Health and Human Services. https://pubmed.ncbi.nlm.nih.gov/37792968/

20.

Olweus

. (1993). Bullying at school: What we know and what we can do. Blackwell.

21.

Ray

McDermott

C. D.

Nicho

. (2024). Cyberbullying on social media: Definitions, prevalence, and impact challenges. Journal of Cybersecurity, 10(1), tyae026.

22.

Sam Altman . (2025). We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made [Post]. X. https://x.com/sama/status/1978129344598827128

23.

Swearer

S. M.

Garcia

A. G.

Damme

A. M.

Palmon

Asay

N. J.

Scalora

M. J

. (2020). Effective bullying prevention and intervention strategies for school professionals. In Worrell

F. C.

Hughes

T. L.

Dixson

D. D.

(Eds.), The Cambridge handbook of applied school psychology (pp. 388–408). Cambridge University Press. https://doi.org/10.1017/9781108235532.023

24.

U.S. Department of Education. (2018). Anti-bullying policies and enumeration: Fact sheet. https://www.govinfo.gov/content/pkg/GOVPUB-HE20_7300-PURL-gpo221082/pdf/GOVPUB-HE20_7300-PURL-gpo221082.pdf

25.

van Aalst

D. A. E.

Huitsing

Veenstra

. (2022). A systematic review on primary school teachers’ characteristics and behaviors in identifying, preventing, and reducing bullying. International Journal of Bullying Prevention, 6, 124–137. https://doi.org/10.1007/s42380-022-00145-7

26.

Zhang

Hong

J. S.

Hao

Lee

N. Y.

Piquero

A. R

. (2022). A latent transition analysis of youth bullying victimization patterns over time and their relations to delinquency. Journal of Interpersonal Violence, 37(7–8), NP5442–NP5470. https://doi.org/10.1177/0886260520958635