Artificial Intelligence as an Approach to Ease Emotions of Email Writers

These kinds of situations are especially emotionally triggering for me, as it is hard to know when someone is being nice versus when someone is being passive aggressive in writing.

As a writing instructor, I care about how students use AI as well as why they use AI for certain tasks. In this article, I take personal experiences as a graduate student and writing instructor and reflect on why and how I either used or could have used AI to meet my professional, academic needs.

Tomkins (1962;1963), the father of affect theory, named nine affects: interest-excitement & enjoyment (the positive affects); surprise-startle (the neutral reset affect); and fear, distress, anger, disgust, contempt, and shame (the negative affects). Email communication is, in my opinion, a risky, affectively negative, and anxiety-inducing communication mode. It is used often in academia, serves as permanent once sent, and opens up email recipients to a wide range of possibilities in topic, audience, and rhetorical situation. It was also a skill that I was not explicitly taught until my fourth year of undergraduate college; teachers may have assumed I’d know how to professionally communicate via email because my generation and the generations afterward grew up with more technology as “digital natives,” which has proved to not directly improve our use of technology for professional communication in previous studies (Kim et al., 2016).     

In addition to growing up with technology, I have other identity factors that make professional email communication riskier for me. I recognize writing as an embodied practice, where one’s identities impact the writing process and communication with others. As a female-presenting academic, I know I may suffer negative consequences of communication errors at a higher level than male counterparts (Frasca, et. al, 2022). I am also neurodivergent, an umbrella term for those with psychological and behavioral differences including autism, ADHD, and mental illnesses among others. Neurodivergence can often impact one’s ability to understand and be understood by others (Kuder et. al, 2023) and lead to othering due to perceived absence of expected emotional affects (Bollmer, 2013).

My neurodivergent communication modes mean I might be more at risk for making social communication errors that others may be less likely to commit or be accused of committing as well. Previous research such as Beck et al. (2024) and Cai & Richdale (2016) correlate low self-esteem and difficulty maintaining jobs with trouble managing emotions and communication challenges in people with autism. Experiences when my neurodivergence negatively impacted my email writing are what inspired this research. While these experiences are directly tied to neurodivergence, the negative affect of email writing can impact neurotypical email-writers, too.

Here, I focus on using AI to relieve users of negative emotions during email composition in five specific situations:

1. to address concerns about timely communication
2. to relieve writers of any compounding stress caused by inbox volume
3. to help writers complete incomplete thoughts through typing
4. to receive feedback on especially unique situations such as email bullying
5. to practice addressing others with different rhetorical choices

Addressing concerns about timely communication

In “Ethics of Pace,” Moya Bailey (2021) draws attention to how a work culture that emphasizes speed creates anxiety and “false disability,” disabling conditions that appear where there wasn’t disability originally. In academia specifically, our students, faculty, and university staff often expect immediate email responses from us. This expectation is grounded in the fact that we are surrounded by technology (smart phones, tablets, laptops, etc.) and because of our instant access to technology, others should also have instant access to us through our technology. This dangerous assumption causes negative effects while using technology to communicate. Lanctot & Duxbury (2022), for example, stress that urgency and importance are not the same and people communicating through email have trouble differentiating between the two.

Additionally, email authors and email recipients have different evaluations of any given email’s importance and urgency based on the email’s topic, author, or connection to one’s job. The disconnect between people in a conversation—albeit a virtual, asynchronous one—can lead to more negative emotions if the sender’s perception of urgency and importance is not met with the expected appropriate response from the recipient. This leaves us with questions about how long is too long between sending an email and getting a response and how long one should wait before sending a reminder email if they have not been responded to.

The first time I used AI outside of teaching training was in a similar situation (Williamson, 2024). When I was frustrated, I drafted an email that angrily accused the recipient of not reading my emails for biased reasoning. I could not let another week go by after my emails weren’t being responded to within three weeks, so I asked ChatGPT to write an email following up on the emails I sent previously. After reading it, I was pleasantly surprised at how ChatGPT remained professional but still addressed my problems in its response. With a couple edits to the organization names and situational information, I sent the email written by ChatGPT. Finally, the email recipient responded and did what I had been asking her to do for almost a month, and she apologized for not responding to my emails or doing what I asked her to do.

Even though the situation was difficult, I agree that using ChatGPT instead of sending an email I drafted earlier revealing all my insecurities was the right choice. Being angry or passive aggressive because someone doesn’t respond to you quickly enough is rude and only worsens the situation. It can be helpful to consider reasons for emails not being responded to soon enough in professional settings. Reasons might include accidental email burying after a large amount of emails, wi-fi difficulties, and clicking the wrong button between common responses like reply and reply all, among others.

Working through this email situation with AI taught me to assume accidental email ignoring instead of intentionally negative behavior and move forward in communication with that assumption in mind by being positive. It included some common moves to improve positivity in the email, such as “I hope this email finds you well,” “follow up,” “I appreciate your time,” and ended of course, with a thank you. None of these important rhetorical moves were in my initial frustrated email.

Inbox Volume Management

In addition to the stress caused by expectations of timely response, email writers can also experience compounding stress from the ever-expanding volume of unread emails in their inboxes. On any given day, I can wake up with over 20 emails in my inbox. Emails include mass #CFPs and #jobpostings from our field’s various ListServs; invitations to participate in research studies; campus-wide and department-wide announcements; announcements from Blackboard or other LMSs; mandatory meeting “invitations” to add to the digital calendar; an email or two from a student who didn’t read the syllabus; an email or two from a student who did read the syllabus and has a legitimate unanswered question; spam from publishers trying to convince me to use their new “groundbreaking” text in my courses; and my personal favorite, the occasional notice of “Free food in the breakroom.” I can only imagine this list of emails getting longer once becoming a full-time lecturer, writing program administrator, or faculty. And just when I finally found time to scroll through all the emails, sorting the junk from the ones I actually need to respond to, I get another email. My inbox is never quite clean.

When the days are filled with class sessions, meetings, and conferences, it’s easy to let the unread emails pile up until the end of the day or the day after that to find hundreds of emails to sift through. The process of going through emails itself could take over an hour. It’s also easy to get intimidated by the amount of time and work it takes to go through the emails. When you’re new to the field or a budding researcher like myself, you may have less experience with the volume and the various importance levels of your emails. Ishii (2005) found that less experienced email users can develop negative feelings about excessive amounts of emails. Though that’s not a big surprise, the emails still need to be dealt with. Under dozens of emails that don’t require your response are a few that really do matter to students, researchers, and co-workers. I personally know that if I don’t respond to truly urgent emails when I first open them, the email will probably get buried under another avalanche of spam, and I might owe someone an apology for not remembering to go back to responding to it.

Some emails require no responses, cookie-cutter language, and more finesse. For those emails that can be answered using language from policies, work documents, or standard procedures, AI can be helpful. AI can respond to emails using a professional tone, careful word choice, and references to important resources and other documents. It can help create templates to use when solving similar issues and discussing similar things multiple times with perhaps multiple audiences.

The goal is to realize which emails this approach would work for and which emails don’t. For example, when I get multiple emails from different students on the same issue with homework, Blackboard, or our textbook, instead of individually replying to all of them, I can create a template announcement using AI to share the information with both students who had questions and students who might have similar questions later on. Class announcements have worked well for me. I can follow the template formatting, greeting, sign off, and content organization for various announcements in class throughout the semester without having to start each one from scratch. I can also input syllabus information for AI to incorporate into an announcement using the template if I’m feeling especially short on time or mental energy.

It’s important to note that AI can pose risks for privacy if it is not used carefully. At my university, copying then pasting student work into AI checkers is a violation of the American FERPA laws because AI can take information inputted as a prompt and use it to train itself, a process called “scraping” (Amarikwa, 2024).

In order to protect the privacy of my students and other people who communicate with me via email, I avoid putting in email addresses, identifying information, and names of students, myself, and the university. I copy email bodies and anonymize any identifying information before submitting it as an AI prompt, and you should too. I also take care to not put anything into AI that I wouldn’t want it to scrape from me. I personally feel comfortable inputting information from my course schedule as the texts I use and my assignments are widely known in my field and tend to be accessible and available outside of my course, department, and university. Especially after training newer graduate assistants with my own teaching materials, I am less concerned about the information from documents being scraped other than the unique course codes and university identifiers. If that’s something you don’t feel comfortable with as an instructor, you get to set those boundaries for when you use AI for mass email announcements or other emails that can benefit from templates.

Help writers complete incomplete thoughts through typing

Information processing is a common challenge for neurodivergent people. It leads to incomplete thoughts or difficulty describing what we know in a way that others, especially neurotypicals, will understand. This is why I personally value having students collaborate and discuss what we read, research, and write in my classes with their peers, writing center consultants, or family members. I had difficulty writing. Talking about it with someone else to help fill in the knowledge and vocabulary gaps was essential for me to transform my thoughts into comprehensible writing.

Ironically, I am more of a night person than a morning person when it comes to having important conversations about philosophy, writing, and theory. Where does one go at 1 a.m. when they need someone to talk through their incomplete, budding thoughts? Now, AI provides us with something that can work like a peer to talk with and share ideas in our own time. It makes it a good option for those who procrastinate until the last minute to write or those night owls who need to work through ideas with someone else when the rest of the world seems to be asleep. What works for class assignments can also work for email ideas.   

AI is a useful way to help figure out how to say what I need effectively and professionally. I can ask it prompts like “what to do when your professor won’t respond to your emails about missing class” or “how to ask for better Blackboard management from your professor in a professional tone.” I’m sure AI would have seen emails I sent in the past and would have told me to explain myself better and teach me more professional ways to rephrase what I meant. It could have also saved me the embarrassment of having to be taught this lesson by a professor when I was a college junior; that shameful story will probably never leave me. AI could fix that with better emails, better interactions, and less negative sense of self.

As stressful situations like these are likely to happen to our students, it’s important to teach them how to collaborate with AI when writing both assignments and other professional communication documents such as emails. While AI can give us a starting point when we struggle finding the right words to express ourselves, we also need to recognize it as just that—a starting point. Just like in-class conversations, conversations with AI happen throughout the writing process, not just at the end. After seeing AI’s responses to our prompts, we have to thoughtfully engage with AI and figure out what works and doesn’t work for each unique rhetorical situation and purpose. We need to revise its ideas to work for our own unique purposes and situations, especially as our emotions as email-writers may vary in type of emotion and how strongly we feel that emotion at any given moment. ChatGPT even reminds users below the chat box that its responses need to be double-checked for any mistakes. These are aspects of AI I teach my students so they know how to use it effectively as a tool rather than it being a replacement for critical thought while writing.

Receive feedback on especially unique situations such as dealing with email bullying  

Some emails are more than unexpected; they’re uncalled for. Receiving an email that uses rude tone or passive aggressive language can elicit a range of negative emotions from the recipient. In addition to consisting of purely assertive or passive aggressive tones, emails can also go further as an act of bullying. Bush and Zuidema (2012) state that email bullying arises when assertive communication seeks to exploit power differences, humiliate or threaten the recipient, or result in feelings of shame, powerlessness, or alienation from peer groups. If it is related to work or comes from someone higher up than you in your organization or institution, the pressure to handle the situation with care and elegance increases. These kinds of situations are especially emotionally triggering for me, as it is hard to know when someone is being nice versus when someone is being passive aggressive in writing. I can think and overthink about an email for hours; if I assume a nice tone over a passive aggressive one, my response would be one way versus if I assume the tone is something else. In the past, I have learned to give myself time to cool down before responding to particularly emotional emails, such as those from students who don’t listen to my feedback. I’ve also seen my peers ask our writing program administrator for advice when confronted with difficult student emails. What is a person to do if the email needs a more urgent response, though?

McCabe (2024) describes Caroline Maguire’s concept of the “social spy” as a benefit for neurodivergent people: “Watching what others typically do (and don’t do) in social situations can help us calibrate our understanding and expectations. You can also study how those with your brain type interact with one another.” However, doing this with email would be a hassle if it is even possible. Would a professor let me watch over their shoulder as they scroll through their piles of emails and figure out how to respond to each one? Would a writing program administrator have examples of testy emails from students, faculty, and staff and how to respond to each? I highly doubt it. I feel like email communication is not given enough attention when new college instructors and graduate assistants are being trained. It’s like we are expected to know this already, while it usually is a learning process via direct experience instead. So how are we supposed to play social spy and learn how to properly respond to especially unique situations via email with grace within an appropriate amount of time?

This is where AI comes in handy. AI can review a received email and write an appropriate response from the perspective of the recipient, which is especially useful in unique, specific rhetorical situations. It’s immediate as well as able to take and apply feedback from the AI prompter (responding in a specific tone, addressing a certain situation, using the perspective of a specific person, etc.). And each time AI writes and revises the email response, the AI prompter and email recipient can learn through reading about a variety of responses they can take toward the specific situation. While learning which response is ready to send and which needs revision before sending takes practice and patience, this way of using AI in email-writing provides much needed emotional relief.

Practice addressing others with different rhetorical choices

Good AI use starts with a good prompt for the AI to respond to. Tone, content, and organization may differ according to email recipient, purpose, and rhetorical situation. Part of what can be anxiety-inducing is not always having a firm understanding of these different audiences, purposes, and rhetorical situations. Can you send the same email to a colleague that you would to a work friend? Or the same greetings for someone you’ve never met compared to your supervisor for five straight years? Probably not. Asking AI to incorporate different audiences, purposes, and situations in your prompt can help you effectively address audiences, purposes, and situations you don’t have as much experience with. It can create templates or highlight different rhetorical moves for you to follow when writing your own emails. You can also prompt AI to address similar topics for different audiences, purposes, and situations and explain why its response looks different for these different situations. This, in my opinion, is a good use of AI because it allows you to act as a social spy without the feeling of invasiveness that comes from observing people in public. The AI does not care if it is being watched or modeled after.

Conclusion

Negative affect I (and I’m sure many other neurodivergent writers or students) experienced while writing emails included fear, anger, and shame. These negative affects have a powerful role in my body. If AI could help manage those negative emotions and turn negative emotions into positive emotions, it could have positive effects on more than just our email writing. It could improve self-esteem, self-worth, understanding of others, and more coping techniques to help during times of stress in the writing process for various genres.

While I am not a big fan of using AI for art due to copyright and artists’ rights as well as concern over environmental impact, I can see the usefulness of AI in contexts specifically related to professional communication in academia and writing instruction. I believe people are turning to AI not just because of those increases in positive affect, but for other reasons I mentioned here including timing, AI as peer for collaboration, and information processing. Email communication is common, important, and risky emotionally. That is why I want to talk to my students about their emotions and state of being when considering using AI for email-writing. Emotions can be roadblocks, but with AI, there are other, better paths forward when writing emails.

References

Amarikwa, M. (2024). Internet Openness at Risk: Generative AI’s Impact on Data Scraping. Richmond Journal of Law & Technology, 30(3).

Bailey, Moya. (2021). The ethics of pace. South Atlantic Quarterly, 120 (2), 285–299. https://doi.org/10.1215/00382876-8916032

Beck, K. B., MacKenzie, K. T., Kumar, T., Breitenfeldt, K. E., Chang, J. C., Conner, C. M., Mandell, D. L., White, S. W., & Mazefsky, C. A. (2024). The world’s really not set up for the Neurodivergent person: Understanding emotion dysregulation from the perspective of autistic adults. Autism in Adulthood. https://doi.org/10.1089/aut.2023.0214

Bollmer, G. D. (2013). Pathologies of affect. Cultural Studies, 28(2), 298-326. https://doi.org/10.1080/09502386.2013.826264

Bush, J., & Zuidema, L. (2012). Professional writing in the English classroom: Are you a writing bully? Considerations for teachers and students. English Journal, 101(6), 100–103.

Cai, R. Y., & Richdale, A. L. (2016). Educational experiences and needs of higher education students with Autism Spectrum Disorder. Journal of autism and developmental disorders, 46(1), 31–41. https://doi.org/10.1007/s10803-015-2535-1

Frasca, T. J., Leskinen, E. A., & Warner, L. R. (2022). Words like weapons: Labeling women as emotional during a disagreement negatively affects the perceived legitimacy of their arguments. Psychology of Women Quarterly, 46(4), 420–437. https://doi.org/10.1177/03616843221123745

Ishii, K. (2005). The human side of the digital divide: Media experience as the border of communication satisfaction with email. Journal of Technical Writing and Communication, 35(4), 385–402. https://doi.org/10.2190/CHEJ-2CW4-6BNY-9LRT

Kim, D.-H., Yoon, H. B., Yoo, D.-M., Lee, S.-M., Jung, H.-Y., Kim, S. J., Shin, J.-S., Lee, S., & Yim, J.-J. (2016). Etiquette for medical students’ email communication with faculty members: a single-institution study. BMC Medical Education, 16(128), 129–129. https://doi.org/10.1186/s12909-016-0628-y

Kuder, S. J., Accardo, A., & Woodruff, J. (2023). College success for students on the Autism Spectrum: A Neurodiversity perspective. Taylor & Francis.

Lanctot, A., & Duxbury, L. (2022). Measurement of perceived importance and urgency of email: An employees’ perspective. Journal of Computer-Mediated Communication, 27(2). https://doi.org/10.1093/jcmc/zmac001

Maguire, C. (2023) Becoming a social spy: How to read people using observational learning. https://carolinemaguireauthor.com/ted-talk/

McCabe, J. (2024). How to ADHD: An insider’s guide to working with your brain (Not against it). Rodale Books.

Tomkins, S. (1962). Affect imagery consciousness: Volume I: The positive affects. Springer Publishing Company.

Tomkins, S. (1963). Affect imagery consciousness: Volume II: The negative affects. Springer Publishing Company.

Williamson, S.J. (2024). “Why I finally used AI.” I Disappoint My Dad Blog. https://sjphd33.wixsite.com/sjphd/post/why-i-finally-used-ai