Content uploaded by Joshua Wilwohl
Author content
All content in this area was uploaded by Joshua Wilwohl on Aug 06, 2024
Content may be subject to copyright.
The Fluidity of Writer Identity with Artificial Intelligence
Joshua Wilwohl
Murray State University
Author Stephen King said about writing: “The scariest moment is always just before you start.”
That time may now no longer seem so frightening when all a person has to do is type a sentence
or two—and let robots handle the rest. Within the past three years, the emergence of new
technology and user-generated content, or UGC, has reshaped identity formation through
writing. The transformation extends beyond individuals and even impacts industries. Pedagogical
practices are pivoting to build upon existing applications, as scholars are urging educators “to
embrace the opportunities presented by this development” (Dergaa et al. 620). Artificial
Intelligence, or AI, may indeed help improve one’s writing, particularly among English
Language Learners, who must “meet expectations placed upon them” (Tseng and Warschauer
259). But reality is that as technology and UGC become more prominent, the definition of
“writer” continues to evolve. The traditional, formal ways of composition are morphing into
informal UGC, blurring the lines between human and non-human authorship. As a result, a
single question arises within writing pedagogy, “Who is a writer?” The answer is murky,
especially when the composer is made of gears and wires, but relies on flesh and blood for a
prompt to awaken the bots. As AI becomes a more active participant in writing and social
networking sites become the platforms of choice, the formalities of text become
messy—potentially reshaping how one understands the self in an era where human and machine
creations intertwine.
1
In its traditional sense, writing records ideas, and through the language used to do so, a
writer can find their identity. This view, though, becomes challenged as the field of composition
moves toward a combined approach, which is no doubt the start of the “new faultlines for
debate” away from “humanist assumptions” about writing (Brooke and Rickert 163). Byron
Hawk discusses language being a “regime of signs” that “operate as variables at the intersection
of a machinic assemblage of human and nonhuman bodies” (79). This intersection appears to be
where identity now forms. While postprocess theory sees a writer’s identity emerge through the
act of writing itself and within social context, looking beyond such ideas with today’s technology
reveals a new narrative—a transference of language composed by bots, but initiated by humans.
The question of identity arises, of course, as the prompt that includes the initial writer’s language
is becoming amalgamated with the language generated by AI. This is important because
language is valuable to experiences. Collin Brooke and Thomas Rickert argue that “language and
technology are constitutive and transformative,” and this cyclical relationship challenges
traditional identities in the sense that they can now be more fluid and less constrained by social
norms (177). This fluidity is highlighted with the emergence of social networking sites, such as
Facebook and X, once known as Twitter.
In Ordinary People and the Media: The Demotic Turn, Graeme Turner discusses two,
three-letter words that define the media industry: old and new. He states of this dichotomy, “The
[old] media studies’ focus upon industries and institutions is displaced by a new media focus on
the ‘produsing’ consumer, customized content and the individualized audience” (Turner 92).
These labels, however, are more than comparisons of media. They are direct connections to one’s
identity, acting as a transformation when it comes to writing. Referencing academic John
Harley’s work, Turner states there is the emergence of a “do-it-yourself or DIY identity” as a
2
result of “a rapidly decontextualized network of meanings which locate identity in the media
sphere” (41). Turner differentiates “old” and “new” by equating “new” with “DIY
[do-it-yourself] projects” (92). The idea that media is now formed in the hands of audiences
instead of companies necessitates a shift in writing instruction. Raúl Sánchez suggests “informal,
unschooled writing practices and pedagogical relationships take place throughout culture” and
that they become “more pervasive and empirically complex in our technology-rich
environments” (189). The environment has never been as rich in technology as it is now, and
such changes have shifted the direction of writing, particularly its relationship with formalities
and identity.
As formalities change, it influences identity as a result of UGC. Similar to AI, UGC often
makes unclear the distinction between formal and informal writing styles. The “new media,” as
referenced by Turner, tends to move the needle in the direction of informality. The traditional
approach, characteristic of old media, has always been a process that highlights specific stages:
pre-writing, drafting, revising, editing, and publishing. This includes the use of specific (usually
prescriptive) grammar and style, often dictated by guides. New media returns control to
audiences for a more unconventional approach that parallels with Katsushika Hokusai’s haiku: “I
write, erase, rewrite / Erase again, and then / A poppy blooms” (Baldwin, “Analysis of ‘A Poppy
Blooms.’”). The economy of language, with a focus on the moment, translates to the writing seen
across social networking sites. Some may view the succinctness and engagement as upsides to
informal writing. Language, though, relates to the platforms or products in which it is posted,
and, as Turner states, “The social networking site is about the construction, maintenance and
performance of cultural identities” (104). Key word being culture, which hints at linguistic
3
diversity—individually and within communities of practice—throughout social channels, helping
identify audiences and tailor writing.
Sánchez offers a radical approach to the relationship between culture, pedagogy, and
writing. He rejects the traditional hierarchy where culture and writing are seen as being
transmitted or taught, and instead focuses on the fact that “writing is culture” and “writing is
pedagogy” (Sánchez 193). Design thought leader John Maeda suggested in a 2012 TED Talk that
when defining “old” and “new” alongside media, differentiating them is not important. Instead, it
is about finding “something in between” by combining the old and the new to discover what is
good (“How Art, Technology and Design Inform Creative Leaders”). While Maeda’s “good”
looked at leadership, his thoughts could help shape an interpretation of Sánchez’s view: writing
as a heterarchy. In this situation, the skill is not a tool or a product, but an active
participant—much like how AI is labeled now. UGC fits into the heterarchy by acting as a
democratic learning space to share knowledge and experiences through content that results in
identity formation. Such an approach, however, requires an overhaul and investment in writing
instruction—one that “enables a customized, endlessly iterative, performance of individual
identity” (Turner 104). This investment creates a narrative of the self, and it is not siloed to
individuals. The impact is also within entire industries, such as journalism, where writing is a
core practice.
Turner predicted in 2010 the decline of the “professional production of journalism,” as
audiences started to become reporters (56). Citizen journalists took to social networking sites to
post what they heard and saw, and their missives became a new form of writing. The use of “new
media,” such as social networking sites, meant there was no sequence, as seen in “old media,”
but instead a narrative of informal writing that defined people. More than a decade later, Dakota
4
Morlan asks bluntly, “Will generative AI replace journalists” (“Journalists Face an Identity Crisis
With AI”)? Not only is writing on social networking sites by ordinary people morphing an
industry, but this is now compounded by the advancements in technology, such as AI, resulting
in murkier questions around identity. Morlan calls it a “crisis,” but such a label seems extreme
under the current circumstances. Humans are crafting AI prompts to generate content, hinting at
some control over identity. The Large Language Models, or LLMs, that AI is trained on also
remain shaped by humans. It is important to note that with this evolution comes possible bias
because of the training data. Humans may still be able to maintain control over the direction of
writing, mitigating biases and a potential shift in identity formation, based on what they type into
the prompt—a nudge “towards unbiased outputs” (MarkovML, “A Comprehensive Guide to
Prompt Engineering”). Journalism schools are attempting to mitigate these practices by
positioning AI as an assistant that helps reduce the workload. While programs, such as
Northwestern University’s “Generative AI in the Newsroom,” provide awareness, they do not
appear to help steer the direction of identity when it comes to writing and AI’s use.
Journalism has techniques and styles that are often associated with formal writing
practices, and journalists (citizen or professional) use their own language when describing
events. It is how they build followers. With the emergence of AI, though, one calls into question
the identity of an “agent”—what Debra Journet in Beyond Postprocess labels the writer (43).
Essayist Evan Puschak argues, “Not to write is to live according to the language of others. Or,
worse, to live through the edits, tweaks, and embellishments to language generated by an
overconfident AI chatbot” (Puschak, “The Real Danger Of ChatGPT”). The irony of Journet’s
term is that the “next big thing” in AI is “agents,” which can “autonomously make decisions in a
dynamic world” (Heikkilä, “What are AI Agents?”). These agents are labeled within the
5
technology industry as systems that help with everyday tasks, from booking hotels to sending
emails. They are also multimodal: “For example, in Google’s Astra demo, users could point a
smartphone camera at things and ask the agent questions. The agent could respond to text, audio,
and video inputs” (Heikkilä, “What are AI Agents?”). It is almost “prophetic” that Journet’s use
of the word “agent” is employed more than a decade later to identify an AI system that
completes our writing tasks (Dickey). Writing itself can be considered a system, functioning as a
form of communication, with its set of symbols, styles, and so-called rules for language—not
dissimilar to how AI agents function with codes and algorithms to process information. Writing
also has input and output mechanisms that requires a systematic understanding of items, such as
the trio mentioned in the previous sentence. AI agents operate the same, but instead of language,
it is data. Journet states there is an “emphasis on agency—both material and semiotic—[that]
reflects increased disciplinary interest in how digital technology and multimodality are changing
what we think of as composing” (55). Technology and multimodality are key because they
introduce new ways to create and to combine within composition. As a result, people can
develop their own identity—“agentive selves”—through the use of tools and cultural materials
available to them at that moment in time (Journet 55). Thirteen years later, Journet’s discussion
remains relevant. Only the tools and culture have changed, but with these changes brings about
an entirely new sense of self.
In 2011, when Beyond Postprocess was published, social media was in its infancy, and
the agent was a theoretical label assigned to a human. Formalities of language started to shift
toward the informal and online identities started to emerge through social networking sites’
profiles. Tools, such as the now-defunct online bookmarking service Delicious, provided a
different platform for writing. Brooke and Rickert suggest Delicious’s tagging feature was a form
6
of writing that “nudges us toward seeing language, world, and human being in fresh light,”
implying how technology is not a passive tool for writing, but is instead an active movement
toward shaping one’s self (177). While the duo argue that “tagging is writing,” the agent at the
time remained human (Brooke and Rickert 177). Now, the five-letter word’s reference is
transitioning from people to bots. With the development of AI, the technology may act as an
extension of tagging by evolving categorization through processes, such as curation. When the
technology becomes more sophisticated and more content is generated, tagging may fully
transition to the role of AI—identifying patterns and blurring the lines between human and
non-human text. This obscurity leads to the questions around identity. Digital platforms now
integrate AI, and some “require human writers to employ AI for automating and analyzing
writing tasks across a range of functions that are becoming ever more tangled” (Pedersen and
Duin 5). This requirement could be seen as a form of nudging toward a “posthuman world,”
where humans are decentralized in the writing process (Hawk 77). While society may help direct
this prodding, it is, ultimately, “a movement towards—towards a text, a position, an identity”
(Journet 41). But nudging is what teachers do to help students, including appealing to their
identities: “[C]hanging behaviour is done by appealing to a student’s current identity, a future
identity (often the later work field), or a different identity altogether” (Weijers et al. 13). This
appeal could be a way of framing lessons in writing, possibly looking at students’ online
identities through social media and comparing them with the authorship of AI. It could also
incorporate multimodal storytelling, which has never been as prominent as it is now. Such a
nudge, of course, raises ethical questions around the use of AI, in particular when it comes to
plagiarism, and its relation to identity. Within the domain of education, students type their
thoughts into a prompt and trust “the software will recommend language to express what he or
7
she is trying to say—cocreating with the human based on his or her ideas” (Pedersen and Duin
4). Co-creation is a common theme in the posthuman world of composition, and it sits alongside
that of identity because they both come down to authorship. Who owns the output? Is it solely
based on the input? In the traditional sense, the writing is owned by the writer. When it comes to
AI, though, it gets hazy.
Humans input ideas into a prompt, and AI generates the content. The initial idea—a few
words, sentences, paragraphs—belongs to the human. The thought is written in the language of
the human, following the grammar and style learned by that person. The prompt becomes
identifiable with them. What is generated after the prompt is submitted, however, loses a sense of
identity as authorship is transferred to the AI—in a language, some may argue, that is almost too
perfect. Aimée Morrison states in her essay, “Meta-Writing: AI and Writing,” that writing to her
feels “allusive and elusive and annoying and impossible” (Morrison 155). All—what King may
say are—scary. But what if, for her essay, Morrison wrote only a 60-word prompt, a broadly
suggestive two sentences engineered to produce 500-plus words about AI and writing? She
attempted a similar experiment for a section of her essay, which showed a stoic response from AI
compared to what she wrote. Of course, manipulating how the prompt is written, such as “in the
style of” or mimicking, could change the output. The point, though, is that there are stark
differences. Morrison’s voice—the enthusiasm—can be heard as her text is read: “When I
encountered the prompt ‘AI and Writing,’ then, I got excited…[L]ike a crackling cloud in a
rumbling sky when the hair on your arms starts to stand up just before the discharge of lightning”
(Morrison 155). The AI response is much more mundane: “Some of the ways in which AI is used
in writing include” (Morrison 156). Morrison’s prose is peppered with bits of character identified
mostly in informal writing, which shapes the reader’s thoughts of her as well as provides some
8
self-reflection—despite Morrison’s self-sabotage of it not being up to snuff for academia. She
raises an interesting point, though, when comparing the two texts, stating there is an
“unaddressed power struggle between human writer and AI content creation” (Morrison 159).
The use of the word “content” for Morrison appears to be reserved mostly for AI, while “writer”
remains mostly in reference to humans. This word reservation may help with furthering identity
formation when lines start to blur: A writer types a prompt to generate content. Along these
lines, words such as “write” vs. “create” and “craft” vs. “generate” may provide hints at who is
writing. Technology companies appear to be at the forefront of upending this idea, though, with
tools, such as the “AI Text Humanizer,” which states, “Paste your AI-generated text and let our
tool humanize it for you. Then copy the results for further use” (Blue, “AI Text Humanizer
Tool”). Blue’s technology adds another layer to the concept of identity when a human writer
crafts a prompt and copies the AI-generated content into a “humanizer,” which then encourages
the human to further use the content. Who is the agent now? There is no clear answer to this
question, but semantics become even more important as society enters the posthuman world.
Most AI-generated content shares a common rigidity in its writing that almost identifies it
as machine rather than human—even with Blue’s technology, which appears mostly to adjust
sentence length with more contractions and pronouns. This AI-guided Standard English may be a
draw for English Language Learners and ideal for students who think they cannot write, but it
may also end up uniforming the lexicon, raising more ethical questions around its use and
relation to identity when everyone’s essays, social networking site posts, and emails start to
sound the same. Instead, using AI as a tool within composition classrooms that supplements
original writing may be the ethical step forward. The focus should be on how AI can help, not on
what AI can do. For example, an AI may assist in developing ideas for an essay, rather than
9
generating the essay. Or it could analyze what has been written to see if the text can be improved
by suggesting or tweaking, rather than creating. The essayist Puschak questions whether society
will move to become a group of editors rather than writers: “Editing, after all, is another way to
synthesize information” (“The Real Danger Of ChatGPT”). Moving in this direction, though,
puts the brakes on a writer’s identity. Journalists see this happen when editors tear through copy,
replacing words and rearranging phrases—a possible reason for migration toward social
networking sites by professional journalists who want to control their own narrative. The idea of
control is the issue as the homogenization of writing styles becomes more apparent. This allows
one to look at identity through the lens of “language as material” and “language as meaning” (de
Freitas and Curinga 250). There is a risk of social networking site posts, essays, and other written
content sounding the same as AI starts replicating the material aspects of language, adhering to
grammatical rules, following specific styles, and mimicking authors’ voices. It can also do the
same when it comes to meaning by analyzing data and identifying how words are used. Some
may argue that this can help streamline information, taking the burden off writing boring copy,
such as routine emails, meeting minutes, data entry, and some product descriptions. The trick in
staying original, though, may be in the nuances of language. AI currently appears to have
difficulty recognizing (1) cultural references, such as shared experiences and double meanings,
(2) figures of speech, such as metaphors and sarcasm, and (3) non-verbal cues, such as tone.
Utilizing these subtleties within the language of meaning may help a writer’s identity emerge in
the era of AI—for now. By contrast, students who are accustomed to UGC and use AI for
content generation appear to “struggle to change the way they write,” depending on the audience
and context (Hartford).
10
Context becomes crucial, and even the ethics of “human-machine writing requires of both
humans and machines a deeper understanding of context” (McKee and Porter 111). Within
writing pedagogy, students should be taught how to craft meaningful expressions because words
shape reality. Or, as Lera Boroditsky argues, cognition: “Studies have shown that changing how
people talk changes how they think” (“How Language Shapes Thought”). A Stanford University
report last year stated an increasing number of academic papers showed “large language models
and chatbots can no longer be differentiated from their human counterparts” (Hancock, “How
Will Chatgpt Change The Way We Think And Work? Stanford Scholar Examines”). The report
raises the question of identifying authenticity in communication, citing a 2023 study that showed
people’s willingness to forgive is hindered if they know an apology is generated using AI. When
a person crafts an apology, which can be considered a meaningful expression, there is a level of
expectation (depending on the situation) that they are writing the message themselves. This
provides a perceived identity of the writer to the reader and shapes the way both think. The
hindrance discussed in the study, no doubt, extends beyond sorry notes to other writings, such as
academic assignments and marketing material, where AI “nudges humans and it affects their
decisions” (Pusztahelyi 21).
As students wade into waters where language becomes indistinguishable between bots
and humans, teachers must find where AI can be used to enhance students’ use of language as
material and can allow them to focus on language as meaning, expressing their identities. The
use of AI tools, such as Grammarly, can help identify errors, while also retaining a writer’s
voice. Of course, too many suggestions could sway the authenticity of the writing. It becomes a
trade-off of what improvements to accept and which ones to reject—all the while human auditors
make changes to the AI based on these decisions, feeding the system bits of people’s identities
11
(Fitria 67). The bots then transform these bits into a combined identity, hence the combined
approach to writing. Such an approach, with the instructional use of AI tools, raises challenges
not only with the identity of the writer, but also that of the writing instructor. Rebecca Moore
Howard argued more than a decade ago that students can perceive instructors as an
“impediment” to their work (226). Being labeled an obstacle—and possibly even
redundant—may now more than ever be on the minds of some teachers, as students take to AI. It
could be that teachers must pivot “beyond” to show students “that their language is valuable, and
their experiences relevant” because the meaningful expressions they write help define them
(Sánchez 193; Morrison 161). It also could be that a balance is the solution: Students use AI to
refine language, while teachers ensure nuances of the human voice resonate through the written
word. It is either one (or both) of those, or everyone turns to the humanizer.
12
Bibliography
Baldwin, Emma. “Analysis of ‘A Poppy Blooms.’” Katsushika Hokusai. Unknown Date.
https://poemanalysis.com/katsushika-hokusai/a-poppy-blooms/.
Blue. “AI Text Humanizer Tool.” Unknown Date.
https://www.blue.cc/resources/tools/ai-text-humanizer.
Boroditsky, Lera. “How Language Shapes Thought.” Scientific American. February 2011.
https://www.scientificamerican.com/article/how-language-shapes-thought.
Brooke, Collin and Thomas Rickert. “Being Delicious: Materialities of Research in a Web 2.0
Application.” Beyond Postprocess. Edited by Sidney Dobrin, J.A. Rice, and Michael
Vastola. Utah State University. 2011. Pages 163-179.
de Freitas, Elizabeth and Matthew X. Curinga. “New Materialist Approaches To The Study Of
Language And Identity: Assembling The Posthuman Subject.” Curriculum Inquiry.
Vol. 45. No. 3. June 2015. Pages 249-265. https://www.jstor.org/stable/43941700.
Dergaa, Ismail et al. “From Human Writing To Artificial Intelligence Generated Text: Examining
The Prospects And Potential Threats Of Chatgpt In Academic Writing.” Biology of Sport.
Vol. 40, No. 2. 2023. Pages 615-622. https://www.academia.edu/103134007/From_
human_writing_to_artificial_intelligence_generated_text_examining_the_prospects
_and_potential_threats_of_ChatGPT_in_academic_writing.
Dickey, April. General Discussion #1. “Historical Principles in Composition Theory.” Paul
Walker. Murray State University. July 1, 2024.
Fitria, Tira Nur. “‘Grammarly’ as AI-powered English Writing Assistant: Students’ Alternative
for English Writing.” Metathesis: Journal of English Language Literature and Teaching.
Vol. 5. No. 1. April 2021. Pages 65-78. https://www.researchgate.net/profile/
13
Tira-Nur-Fitria/publication/352207965_Grammarly_as_AI-powered_English_Writing_
Assistant_Students'_Alternative_for_English_Writing/links/60bed29992851c
b13d88d0eb/Grammarly-as-AI-powered-English-Writing-Assistant-
Students-Alternative-for-English-Writing.pdf.
Hancock, Jeff. “How Will Chatgpt Change The Way We Think And Work? Stanford Scholar
Examines.” Stanford Report. Stanford University. Feb. 13, 2023.
https://news.stanford.edu/stories/2023/02/will-chatgpt-change-way-think-work.
Hartford, Nicole. General Discussion #2. “Historical Principles in Composition Theory.” Paul
Walker. Murray State University. July 11, 2024.
Hawk, Byron. “Reassembling Postprocess: Toward a Posthuman Theory of Public Rhetoric.”
Beyond Postprocess. Edited by Sidney Dobrin, J.A. Rice, and Michael Vastola. Utah
State University. 2011. Pages 75-93.
Heikkilä, Melissa. “What are AI agents?” MIT Technology Review. July 5, 2024.
https://www.technologyreview.com/2024/07/05/1094711/what-are-ai-agents/.
Howard, Rebecca Moore. “Postpedagogical Reflections on Plagiarism and Capital.” Beyond
Postprocess. Edited by Sidney Dobrin, J.A. Rice, and Michael Vastola. Utah State
University. 2011. Pages 219-231.
Journet, Debra. “What Constitutes a Good Story? Narrative Knowledge in Process, Postprocess,
and Post-Postprocess Composition Research.” Beyond Postprocess. Edited by Sidney
Dobrin, J.A. Rice, and Michael Vastola. Utah State University. 2011. Pages 41-60.
Maeda, John. “How Art, Technology and Design Inform Creative Leaders.” TED Talk. YouTube.
Oct. 10, 2012. https://www.youtube.com/watch?v=WAuDCOl9qrk.
14
MarkovML. “A Comprehensive Guide to Prompt Engineering.” Jan. 4, 2024.
https://www.markovml.com/blog/prompt-engineering.
McKee, Heidi A. and James E. Porter. “Ethics for AI Writing: The Importance of Rhetorical
Context.” AIES. Feb. 7-8, 2020. https://dl.acm.org/doi/10.1145/3375627.3375811.
Morlan, Dakota. “Journalists Face an Identity Crisis With AI.” Comstock’s Magazine. Aug. 29,
2023. https://www.comstocksmag.com/web-only/journalists-face-identity-crisis-ai.
Morrison, Aimée. “Meta-Writing: AI and Writing.” Composition Studies. Vo. 51, No. 1. 2023.
Pages 155–161. https://rws511.pbworks.com/w/file/fetch/154174671/
Morrison%20Meta%20Writing.pdf.
Pedersen, Isabel and Ann Hill Duin. “AI Agents, Humans and Untangling the Marketing of
Artificial Intelligence in Learning Environments.” Hawaii International Conference on
System Sciences. 2022. Pages 4-13. https://scholarspace.manoa.hawaii.edu/
server/api/core/bitstreams/522ac070-0c6a-4807-93c4-ba7d2d7f3d8a/content.
Pusztahelyi, Réka. “Emotional AI And Its Challenges in the Viewpoint of Online Marketing.”
July 2020. https://www.researchgate.net/publication/343017664_EMOTIONAL_AI_
AND_ITS_CHALLENGES_IN_THE_VIEWPOINT_OF_ONLINE_MARKETING.
Puschak, Evan. “The Real Danger Of ChatGPT.” Nerdwriter1. YouTube. Dec. 30, 2022.
https://www.youtube.com/watch?v=AAwbvGywdOc.
Sánchez, Raúl. “What Constitutes a Good Story? Narrative Knowledge in Process, Postprocess,
and Post-Postprocess Composition Research.” Beyond Postprocess. Edited by Sidney
Dobrin, J.A. Rice, and Michael Vastola. Utah State University. 2011. Pages 183-194.
15
Tseng, Waverly and Mark Warschauer. “AI-Writing Tools In Education: If You Can’t Beat Them,
Join Them.” Journal of China Computer-Assisted Language Learning. Vol. 3, No. 2.
2023. Page 258-262. https://doi.org/10.1515/jccall-2023-0008.
Turner, Graeme. Ordinary People and the Media: The Demotic Turn, SAGE Publications,
Limited, 2009. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/
murraystate/detail.action?docID=483413.
Weijers, Robert et al. “How Do Teachers In Vocational And Higher Education Nudge Their
Students? A Qualitative Study.” European Journal of Higher Education. Feb. 12, 2024.
Pages 1-19. https://doi.org/10.1080/21568235.2024.2319087.
16
