Dana LeTriece Calhoun
University of Pittsburgh
This assignment invites undergraduate students to create "spells," a poem based on a wish, and prompt an LLM to write a spell of its own, then reflect on their input and LLM output compared to their original compositions. Students analyzed the function of language and intent in manual and LLM composition, drawing together meaning and expression, and how LLM technology replicates or revises that expression. This assignment can be adapted to most rhetoric, composition, and cultural studies courses.
Original Assignment Context: End of elective Public and Professional Writing course
Materials Needed: An accessible AI text generation program (i.e. ChatGPT)
Time Frame: ~2-3 weeks
This assignment, “Spellcraft & Translation: Conjuring with AI” is a part of the Black Rhetoric and Public Writing course taught at the University of Pittsburgh in Spring 2023. This course is situated within the Public and Professional Writing concentration and Diversity in General Education requirement for the university curriculum.
This assignment is placed alongside theoretical and practical work discussing the intersections of Black rhetoric and language with digital hybrid models of analysis and composition. Conjuring hails from Hoodoo, a Black Southern American spiritual and ontological practice first envisioned by enslaved people and manifesting now across a multitude of geographic locations, experience, and identity. In this course, Conjuring acts as the methodological model to ask students to blend the natural, the self, the culture, and technology in scholarly analysis and composition in Black rhetorics. I have taught this assignment once and plan to teach it again in similar courses in the future.
Spellcraft & Translation asks students to compose a comparative analysis between two originally-composed “spells” of their own and spells composed by the AI writing platform ChatGPT. Spells, in this case, are defined as technical instructional documents that convey an intention, purpose, and direction. The assignment before asked students to create a “Mojo Bag,” a repository of tools and objects central to the creator that enhance their work. In traditional Conjuring practice, Mojo Bags are mobile “toolkits” holding personalized items and key ingredients for their work. This assignment was also produced digitally—I wanted to give students the opportunity to display abstract ingredients, like the overall “vibe” of a song that would help them conjure their work. In their bags, students included a variety of tools and objects and described their meaning: a dice, to help them come to a decision about what to write; moss agate, to encourage emotional stability; a poker chip, to represent the act of weighing multiple decisions; and a playlist of songs that encourage self-love, confidence, and power.
ChatGPT is an online AI writing platform that can compose text based upon prompts entered by the user, although any similar generative AI platform would work for this assignment. For this assignment, I was curious about the capability of ChatGPT to generate a text with similar theme, feeling, and composition as students had. From the “Mojo Bag” assignment, I already knew what personal themes or wishes would manifest in their spells: general anxiety about being an undergraduate student, balancing responsibility and relationships, and finding their identity and individuality. In previous classes, I have used spells and poems as a teaching text, and encouraged students to create their own should they feel motivated to as a writing exercise. The idea for using ChatGPT to compose a spell came from a desire to connect Black rhetoric, genre writing like poetry and recipes, and AI writing technology.
The inspiration for Spellcraft & Translation came from my mentor and colleague, Annette Vee, encouraging me to look at LLM technology in relation to my personal scholarly work with Technogothic Conjuring, a methodology I developed while writing my dissertation. Technogothic Conjuring is a methodology taught by Black Southern Hoodoo spiritual and cultural practices and made unique and deeply personal by the blending of digital compositional and archival methods that focus on the identity of the subject and researcher. Conjuring hails from the African American “embodied historical memory” tradition of Hoodoo. Other names include: rootwork, black magic, witchcraft, devil’s work, superstition, spirit work, and conjure.
The “Technogoth” in my methodological phrase deals with the approach towards composition while acting as a descriptor for the practice of blending the technological with the intense flood of emotions that embody the “gothic.” More specifically, the Technogothic approach explores the combination of digital composition and communing with the dead. The Technogothic approach is housed in autoethnography, in that the researcher and the research itself are not separate, but placed in different planes of medium, physicality, and temporality—the “Conjuring.”
My dissertation involved transcribing six months of my great-grandmother’s handwritten everyday journals, and conjuring her spirit and memories by displaying her history on a website. All of the transcriptions and data were done by hand, and one chapter is entirely composed of original poetic remixes of my great-grandmother’s journal entries. After doing all this work myself, I wondered how LLM technology would influence the process of collecting and composing the sheer amount of data available and perhaps inspire a new form of output.
I’ve often discussed and encouraged poetic writing in my previous courses, and students have responded positively, saying that creative writing forms allow them to express emotions, deep thoughts, and opinions more freely and that doing so helps them begin to draft more traditional writing forms. With the Black Rhetoric course Spellcraft & Translation was housed in, I was curious to see how student composition compared to LLM compositions and how students would react and reflect on this process, especially considering the boost in discourse about LLM technology in academia.
I feel that this assignment is able to be transformed and conjured in its own way to fit the needs of a specific course or module simply because AI writing platforms provide a third-party composer that acts as an amalgamation of all types of writing. As a pedagogical tool, AI writing platforms can act as a “shadow writer,” creating examples to analyze in real-time, to either highlight the technical or contextual elements.
Students displayed their comparative analysis on a digital Notion page, including a descriptive analysis of their process of composing their spells and prompting ChatGPT to compose spells of its own and a reflection of said process. Notion is a digital workspace that I’ve found to be incredibly useful for planning and composing digital projects, but any online space that allows students to display their work in different media would suffice. Platforms I have used in the past and encourage students to use have been: WordPress, Wix, Tumblr, and GitHub. Alternative platforms for short-form writing or writing that depends on hyperlinked context include Twitter and Medium.
The purpose of this assignment is to introduce multiple rhetorical concepts such as genre, comparative analysis, poetics, instructional technical writing, knowledge and skill in working with AI writing technology, and sociocultural contexts. The scaffolding for this assignment included students engaging with external readings (included in the assignment description) through a workshop and discussion-based format where students brought in their materials and as a class we discussed their wider implications.
This assignment could be accomplished physically, textually, or digitally, with this specific iteration being the latter using Notion.
The majority of the preparation and materials for this assignment are digitally involved, including knowing how to use ChatGPT and knowledge on how to create effective prompts. This assignment will also involve an understanding how Notion works to display digital comparative analysis projects as well, if presented as a digital project.
Theoretically, the preparation for this assignment involves readings surrounding Black rhetoric and language; digital composition and AI writing technology; topics of race, gender, and technology; and how aspects of Conjuring relate to critical and scholarly composition.
I can imagine, and deeply encourage, revisions and remixes of the main idea behind this assignment—a comparative analysis between two texts. The inclusion of ChatGPT as an interface that students can use to “conjure” a text is an addition of my own curiosity and excitement of the promise AI writing tools have in the classroom.
At first, my students were cautiously excited about working with ChatGPT, having seen numerous reports about cultural anxiety surrounding AI writing. At first we played around with developing silly prompts such as “Write a haiku in the style of Eminem.” We then moved forward to more applicable prompts for this assignment
In our discussion of sourcing the ingredients for the spell, looking deeply at literary and contemporary examples of rhetorical spellwork, including select readings from non-fiction examples like Memory Dishes: Women and African Diasporic Cooking and fictional examples from Beloved by Toni Morrison. We discussed the elements listed in each spell, like chamomile, lavender, and the color pink in Beloved. In class we went through a research exercise finding the traditional applications of these ingredients. I wanted to make the connection between recipes and spellwork through these exercises, to call attention to rhetorical form and construction while situating this form of composition in a sociocultural context. Students immediately understood the connection, and quickly introduced more contemporary examples like the mobile phone acting as a multi-tool or knitting needles to weave things together.
Working with ChatGPT was an exercise in understanding genre, rhetorical emphasis, and the limitations of AI composition. On a pedagogical note, this exercise was a joy to do and brought fun into the class as well as understanding. We began by opening OpenAI’s ChatGPT and prompting it to compose summaries of popular films. Students offered to prompt ChatGPT with more specific language, such as including instructions about using “formal” or “slang” language depending on the summary. We explored the limitations and collated imagination that ChatGPT was able to produce, though the general conclusion was that the AI-composed writing was too “stilted” or “obviously not organic.” The students reasoned that while most of the AI compositions made sense, they failed to see a “spark” within the writing that made it rhetorically effective.
I was delighted at this response about not seeing a “spark,” and we discussed how knowing the source of the output influenced their observations and conclusions. Students brought up a great point that I will add to future iterations of this assignment, that I should provide multiple examples of a piece of writing and ask future students to see if they can choose which was written by a LLM.
Moving towards working with spells and ChatGPT, we prompted it to create spells to obtain good grades, find a new friend, or pass an exam. This exercise was conducted in class, and students were encouraged to play around with prompts and then share their output to a display screen so we could discuss ChatGPT’s output. We continued in a workshop format, where students would analyze and offer suggestions on how to refine the prompt for a more accurate ChatGPT output by including an additional specification, like “include candles,” or “restrict flame.” Together we noted the differences in ChatGPT’s output of these spell prompts, focusing our attention on ChatGPT’s revisions or lack thereof.
Students were particularly amused by ChatGPT’s statement of “As an AI language model, I cannot create real spells or encourage the practice of magic,” which prompted discussion of the human labor that went behind creating this type of technology and this statement being an acknowledgement of the idea that a “real” spell cannot be artificially created.
Despite ChatGPT’s disclaimer, it still composed a spell offering a list of ingredients and the process, though it became evident that most of the ingredients listed were the same across all example spells, along with the process. ChatGPT would offer reasoning behind its choice of ingredients with a generalized summary of the purpose. For example, ChatGPT states that “a green candle might be used for attracting abundance, while a pink candle might be used for attracting love.” Class discussion shifted into researching the historical and cultural meaning behind colors and associations with candles and rituals. We ended this session by calling back to the literary and contemporary spell examples, imagining and drafting what we imagined ChatGPT might create using these examples. The following class sessions were focused on drafting this assignment, working through questions, and further exploration of ChatGPT.
Student 1 prompted ChatGPT for instructions on how to create a spell, which resulted in four steps: determine the purpose of the spell, choose the language of the spell, research the method and symbolism of the words, write out the spell in the chosen language, and practice speaking the spell aloud.
Student 1 followed ChatGPT’s advice in creating their spell “Memoria Della Vittoria,” meaning “In memory of the victories” in Italian. The wish is for the Italian national soccer team to win the World Cup. The spell itself calls for a cauldron, a hair of the last goal-scorer of the Italian national team, a pinch of grass from a sacred soccer field, an Italian national team jersey, and a tear shed from not qualifying two years ago. All ingredients are to be mixed in the cauldron, where the fumes will disperse into the air. Then chant “Memoria Della Vittoria” three times and wait for the wish to come true.
Student 1’s reflection to this assignment addressed the composition of “Memoria Della Vittoria,” citing their Italian heritage and enthusiasm for Italy’s soccer team to win the World Cup. The wish itself may seem simple, but the specificity in its ingredient belies a greater depth. Each ingredient is timeless, hinting toward the history of the Italian national soccer team and its fans.
This example showcases the act of conjuration not only in its content but in the relationship between this writing, this student, and the wider world. I have the privilege of being able to know what the Italian national soccer team means to this student, because they wear the team’s jerseys in class almost every week. They sneak peeks at their phone to check the match scores during our writing breaks. They engage in class conversations making connections between their identity and rhetoric and writing. This spell combines writing and passion into an act of acknowledgement and engagement. It is a blend of the natural, the self, the culture, and technology, fully in line with the methodological approach to this course.
Student 1 prompted ChatGPT to “create a spell,” and the response was to create a spell for positivity and good vibes. ChatGPT’s spell included a list of ingredients, including a white candle, a small glass jar, a handful of dried lavender flowers, a piece of paper and a pen, and a lighter or matches. ChatGPT's spell then included the process for completing this spell. The process was to set up the ingredients and light the candle, then to take the piece of paper and write down the purpose of the spell and to “use positive language.” Then it instructs to fold the piece of paper and place it in the glass jar, sprinkle the dried lavender flowers on top while visualizing being surrounded by a “peaceful aura of light,” then say:
Spirit of the universe, hear my plea,
Bring me the positivity I need.
Fill my life with good vibes and grace,
And keep me in a peaceful and happy space.
ChatGPT then instructs to light the paper on fire using the candle, but advises caution and suggests using a fire-safe container if necessary, then to blow out the candle and place the glass in a prominent place. The conclusion of ChatGPT’s spell is a disclaimer that “this spell is just for fun and entertainment purposes, and it is not intended to replace action or efforts to improve your life.”
During class sessions experimenting with ChatGPT, students often remarked on the disclaimers included in ChatGPT’s output, like the warnings about safety and spells being for entertainment purposes mentioned above. I recalled previous class discussions about authenticity in language and writing, and how Conjuring’s history includes deep discussions about Conjuring fraud that could lead to negative results or injury due to ignorance about ingredients or process. Students connected authenticity and responsibility with ChatGPT’s disclaimers, and wondered if these disclaimer statements or statements of safety were manually included by ChatGPT developers in order to mitigate responsibility.
This student described wanting a “straightforward approach” to create a spell, and questioned ChatGPT’s ability to read their browser’s history, saying that ChatGPT’s spell could have been influenced by their past search history. ChatGPT does not have the access or ability to read a user's browser history, but I find this suspicion fascinating. As an alternative version of this project, instructors could discuss privacy and data settings, surveillance technologies, or power through technological oppression.
Student 1 also questioned the language that ChatGPT used, remarking about its formality. This student wondered “If I use very informal language with it, will it emulate how I speak?”
During a session where students and I were prompting ChatGPT, they requested to see ChatGPT’s output when prompted to give a summary of the movie Shrek with various variations on the form of the summary and the type of language used, ending with “Write a summary of Shrek, use slang.”
ChatGPT’s output was a disjointed summary of Shrek beginning with, “Yo, lemme tell ya ‘bout dis flick called Shrek.” The students were shocked, I was amused, and together we discussed how ChatGPT might have come up with this input. Since this course focused on Black Rhetoric, students spoke to our readings for this assignment. They noted the specific dialect used in Toni Morrison and Alice Walker’s novels, particularly in how they highlight features of African American Vernacular English (AAVE, another term used is African American English).
In classes before, we discussed how Morrison and Walker’s use of dialect in their works is a rhetorical move, one that challenges standardized language use. During this discussion, students connected standardized language use to formality and informality in writing and speech. That conversation led to a discussion of how formality is established—”Clothes, signing emails professionally, not using slang” and how Morrison and Walker’s work challenges the idea of formality in speech and writing.
I guided the conversation towards their prompt of slang, asking them where some current slang terms come from. They referred to fashionable or stylish choice as “drip,” and “snatched,” which is when someone’s appearance is perfect—so good you snatch your wig off, a term and phrase popularized by the Black drag community and further pushed into the cultural lexicon by television show Ru Paul’s Drag Race. Students brought up the history of “snatched” without any prompting, and they named TikTok as a popular source for their current language. TikTok is a social media platform where users upload video clips, and students noted that their “personalized” TikTok feeds featured creators offering commentary on popular culture, Ru Paul’s Drag Race being one of them. Because of these videos, students were mostly already aware of the Black queer and drag community’s historical influence on a term like “snatched.” We discussed how language shifts and evolves from communities into popular usage and how that relates to language appropriation.
After ChatGPT’s output, students connected Walker and Morrison’s dialect with slang terms and appropriation, saying that it was “creepy” and “strange” that ChatGPT affected a similar dialect to Walker and Morrison’s writing. They brought up how contemporary celebrities like Ariana Grande and Awkwafina adopted a “blaccent,” an imitation of AAVE without the racial and cultural background as a form of cultural appropriation.
We then had a brief exercise where we listed common slang phrases and followed the etymological thread back to the communities that popularized these terms in order to create a sample of the terms a hypothetical AI writing platform would encounter if it was trained using popular sources. In most cases, common slang terms originated or became popular from Black and queer communities. Students discussed the possible sources that ChatGPT was trained with, and discussed how they were uncomfortable with the transparency between slang and appropriation.
(A small note: Student 2 used Notion AI, a new feature that debuted during the composition of this assignment—I had not anticipated Notion developing their own AI writing tool, but I encouraged all avenues of AI writing engagement for this assignment.)
Student 2 composed their spells in a poetic format, combining listing ingredients and process in verse form. This student prompted Notion AI to create a spell based off of the wish to not overthink and take things too personally. Student 2’s spell begins with lingering “in cloud of fog,” and then calling upon the wind, “To clear out the disarray,/ And summon the sun to appear and alas,/ Cast the prominent clouds away.” This spell manifests their ingredients and process in a more metaphysical, spiritual, supernatural and natural way, calling upon the speaker’s inherent powers to clear the weather to reflect the clearing of their negative thoughts.
Notion AI’s version of this spell begins with, “By the power of the universe and the strength of my will,/ I call upon the energies of calmness and still./ With this spell, I ask for the ability to let go,/ Of thoughts that bring me low.” This spell continues in a declarative way, ending with, “So mote it be,” a closing phrase used to end rituals or special rites.
Student 2 immediately noticed the formulaic composition of the AI model, describing them as “adjusting their wording and phrasing to incorporate the main asks of the wish,” noting that the formality in the AI spell makes finding the original source easier. Student 2 continued, saying that the AI spell felt like an “expected result,” though not particularly “robotic and lifeless,” which I feel adds to their observation of ChatGPT drawing from external sources that use the same poetic format and closing phrase “So mote it be,” commonly used in Wiccan prayers, spells, and rituals to mean “So shall it be.” ChatGPT’s output here adheres to a type of standardized format for a spell, and though the words pulled from other sources are emotional and evocative, the construction itself belies its uncanny composition.
Student 2 outlined their process and intention in creating their spell, saying that they chose a more figurative and illustrative approach to “create a better sense of vulnerability…and feed into the magic behind them.”
Goals and Outcomes
The learning goals and outcomes for this assignment were to analyze the function of language and intent in manual and AI composition, understanding the implications of the social and cultural impact on AI and similar digital technological composition, and experience in composing and crafting a digital project relating to Black rhetoric and digital writing.
Conjuring with AI is an assignment with a multitude of possibilities, especially with the rapid development of AI writing technologies. Though only two student examples were discussed here, I found that both held common elements and conclusions across the classroom—that composition and rhetorical exploration hold space for magic in the way of imagination, intention, and action. We were able to have productive conversations about the relationships between technology and language, and specifically how AI writing technology displays that relationship.
Most students, in some way or another, addressed the uncanny lexicon of ChatGPT and Notion AI and described that though the language was adequate in achieving the form and general idea behind a spell, the results lacked an element of individuality and humanity. Surprisingly, students also were impressed at how ChatGPT and Notion AI were successful at helping them draft their reflections, providing a general outline of ideas and how to tie them together or provided a quick and clear answer to a clarifying question about an unfamiliar word or phrase the AI used.
I find the generative possibility of LLM technology inspiring as an opportunity to offer students a “playground” to test out prompts, lines of argument, and genre styles. I feel with guidance and support, LLM technology could reduce writing anxiety and writing block in students by allowing them to experiment with a flexible writing utility that can offer questions, generate a general outline, or even produce an example of a paragraph with the style and organization prompted by the user.
Conjuring with AI was an experimental foray into weaving AI writing technology, rhetoric, and composition together. Overall, the resulting work from my students has given me more inspiration to continue to expand both the theoretical and practical application of this assignment. I’m especially interested in adding more readings that address writing, technology, race, and gender into the scaffolding to highlight human identity and labor that contributes to rhetoric and technology.
In future iterations of this assignment, I hope to expand on using AI writing platforms as a point of comparative analysis and compositional tool and introducing it earlier in the scaffolding of this assignment. I found that exploration and experimentation with AI writing platforms tended to be the hot topic of class discussion. I feel that if students became more comfortable with AI writing platforms, their focus could shift more to the rhetorical implications and practical applications, pushing us to consider the possibilities of this technology. I thoroughly enjoyed this assignment, and I am excited about its potential.
I have a bouquet of flowers to give as thanks and recognition: Annette Vee, who has inspired this project by introducing me to ChatGPT and the amazing work that can be done with it along with being a patient and extraordinary mentor; Geoff Glover, for always finding the time to sit and riff, and who always meets my nontraditional ideas with excitement and support; and a trio of perennial blooms for my great-grandmother Mary Ellen Petty, who has walked with me through my scholarship; Laura Mae Freeman, my grandmother who is like my second author with her wisdom etched in my words; and my mother, Theressa Michelle Logan, who always said I would be a shining star. None of my work would be anywhere without you all. Thank you.
Genre is a set of classifications that define consistent elements shared by multiple texts, inside and outside of traditional literary forms. This assignment will ask you to explore the genre of a spell, compose and conjure two spells of your own, prompt an AI writing program to write its own spell, and then perform a comparative analysis between the two texts. Finally, you will reflect on this process of spell creation, AI conjuring, and how writing is shaped with intent and genre.
All components of this assignment, including your spells, ChatGPT’s spells, and your reflection should be hosted on your personal Notion site under the link “Spellcraft & Translation.”
Spells are a specific genre that blend together intent, poetics, rhetoric, and technical writing. One of your spells should focus around your “wish” (a declarative statement that describes a desire you have). The focus of your second spell is up to you! Compose your spell including a list of ingredients and instructions for the implementation of your spell. Your ingredients and instructions may be metaphorical, figurative, creative, or concrete.
After you have constructed your spells, use the skills developed in class sessions to prompt ChatGPT to conjure two spells of its own. Then, compare and contrast your spells and ChatGPT’s spells. In this comparison, you should analyze each spell in depth, discuss the possible purpose and meaning behind the spells created by ChatGPT in comparison to your own, and explore how ChatGPT’s spells compare to your own. Below are questions that I would like you to consider in your analysis:
As always, reflective decompression for your project can be a way to take a step back and take in the entirety of your labor. I would like you to reflect on your process in creating your spells. Below are the questions I would like you to answer, along with any additional thoughts you have:
Create a new Notion page titled “Spellcraft & Translation.” Display both of your original spells and the spells created by ChatGPT. When displaying the spells created by ChatGPT, include the series of prompts that you used.
Introduction to Genre (via music)
Poetry and poetics as genre
To begin the scaffolding process for this assignment, I think it’s important to start with an overview of genre. In this iteration of this assignment, I began with situating the discussion of genre within music. I found this to be an effective starting point, as it encouraged students to offer their own examples of different music genres. We worked through their examples together, finding and notating the similar elements that make up a musical genre.
Then, we shifted focus to poetry and poetics, which I find to be one of the closest literary genres to a spell. Below are examples of poems used as examples in class to establish the genre, explore the purpose of intention and form, and provide a basis for analysis and comparison as practice for the larger assignment.
Recipes as Genre
Next, I introduced the recipe as a genre, as it is very similar to traditional spells in terms of form with a list of ingredients and a description of the process. During this time, we discussed the elements that create the genre of the recipe. In the examples below, we discussed the historical context behind recipes in association with Black Rhetorics, and how this form holds more purpose and meaning aside from being a set of instructions to create a dish.
I asked students to find examples of recipes, and together we contrasted the writing in class to pull out differences in form, language, description, and instruction. Using these recipes, we discussed how to recognize and analyze syntax, diction, and other technical elements of this type of writing.
Spells as Genre, Setting Intention, Sourcing Ingredients
Below are examples of spells that range from literary text to a musical spell that aims to bridge the discussion between genre and spellwork. These examples offer a concrete version of a spell that students can analyze, understand, and possibly gain inspiration from. As a class, we have mapped out each of these spells in terms of intention, form, and performance. With these spell examples, we also defined how each has ingredients (whether physical or figurative) and how they are implemented in the instructions of each spell.
Introduction to ChatGPT
Finally, we began to explore ChatGPT in terms of its creation, function, and uses. Below are articles describing ChatGPT and how other people have used them and its cultural and social impact. Alongside this, the class created prompts for ChatGPT and discussed how the software interpreted them in its output. By creating prompts together, we worked through finding a method for purposefully communicating with ChatGPT, especially in terms of tone (formal, informal, verse, etc.).
Baraka, Amiri. “Babylon Revisited.” Poetry Foundation, https://www.poetryfoundation.org/poems/42559/babylon-revisited.
Bogost, Ian. “ChatGPT Is Dumber Than You Think.” The Atlantic, 7 Dec. 2022, https://www.theatlantic.com/technology/archive/2022/12/chatgpt-openai-artificial-intelligence-writing-ethics/672386/.
Brown, Ariana. “For Everyone Who Tried on the Slipper before Cinderella.” Poets.Org, https://poets.org/poem/everyone-who-tried-slipper-cinderella.
Collings Eves, Rosalyn. “A Recipe for Remembrance: Memory and Identity in African-American Women’s Cookbooks.” Rhetoric Review, vol. 24, no. 3, Nov. 2009, pp. 280–97, https://doi.org/10.1207/s15327981rr2403_3.
Collins, Barry. “ChatGPT: The Weirdest Things People Ask AI To Solve.” Forbes, 4 Mar. 2023, https://www.forbes.com/sites/barrycollins/2023/03/04/chatgpt-the-weirdest-things-people-ask-ai-to-solve/?sh=742f6d8fe547.
Giovanni, Nikki. “Dreams.” Poetry Foundation, https://www.poetryfoundation.org/poems/48224/dreams-56d229494e255.
“Glossary of Poetic Terms.” Poetry Foundation, https://www.poetryfoundation.org/learn/glossary-terms/genre.
Kittelstad, Kit. “Examples of Poetry Genres: Major Styles Explained.” Your Dictionary, 9 July 2021, https://examples.yourdictionary.com/examples-of-poems.html.
Memory Dishes: Women and African Diasporic Cooking. Brown University, 29 May 2019, https://simmonscenter.brown.edu/sites/default/files/pdfs/xMemoryDishes_DRAFT_190514.pdf.
Sherman, Charlotte. Killing Color. 1st edition, Calyx Books, 1991.
Smith, Danez. Dear White America. https://www.poetryfoundation.org/poems/150542/dear-white-america.
Snow, Olivia. “‘Magic Avatar’ App Lensa Generated Nudes From My Childhood Photos.” Forbes, 7 Dec. 2022, https://www.wired.com/story/lensa-artificial-intelligence-csem/.
Walker, Alice. The Color Purple. Harcourt, 1992.
Zeal & Ardor. Devil Is Fine. mp3, MVKA Music Ltd, 16 Dec. 2016.