Liv Owens

initial thoughts + framing

I want to start with the note that this is my first blog post—I have been inspired by my colleague Lily Bichard Collins who has been consistently writing her own blog about her experiences in a PhD program, what writer’s block and stage fright looks like in academia, and general thoughts on the world around us from the lens of ~sociological imagination~. On top of analysis that I admire and a writing style I find accessible and engaging, her blog has been inspiring me to get back to the process of writing for the sake of disentangling and documenting. The further I sink into academia, the more I shy away from the act of writing, ironically. I can get lost in the need to have perfectly articulated, theoretically rigorous, and cutting-edge discourse ready to go at all times. That can force me into a mindset of an all or nothing approach that paralyzes me from writing at all. I want to be brave and consistent like her in my practice and release the need for perfectionism. I just want to write again! I want to have variety in my work and write for the sake of expression and to encourage my own self-reflection.

So, in this vein, I am attempting to block out some time for smaller pieces that help me (and hopefully others) look at the world we exist and participate in and reflect. The goal for these blogs is just that. In a time of constant AI slop, I need this for my brain more than ever. I am flagging now that this post, and the subsequent ones that follow, will be written by an imperfect human. This is going to be a more informal, discursive style of writing to help me think through topic areas that are on my mind. Sometimes it will be cited by “institutionally approved” sources, sometimes I will link to Reddit or a poem or a post that stuck out to me. It will be filled with a mix of American and British English spelling, and we are just gonna let that be. The point is to think think think, and to untangle all those thoughts, and hopefully others will enjoy reading them.

the “open secret” of academic AI use

It has been weighing on me recently that most of the writing I come across now—almost entirely on the internet—is written by LLMs. The great outsourcing of our cognitive abilities to machines is one that is well-documented, this is no means a unique observation that I have made.

For example, the AI detection startup Originality AI found that more than 54% of English posts over 100 words on LinkedIn are AI-written. This is not surprising, as someone who is required to use the platform to network + update colleagues about conferences and funding opportunities in our circles. As my friend Ollie said recently:

“I keep seeing people taking snippets of deeply personal experiences and hollowing out the substance of those moments into a set of points they ‘learnt’ from. It is the classic, ‘my wife left me…and here’s what it taught me about B2B sales’ type post on LinkedIn now. It is as if the experience only becomes meaningful once processed [by a machine] into a neat, corporatized learning metric.”

Working on and in a university institution now, I am struck by how many people use AI for their writing and editing. In an industry that is supposedly about critical thinking and original contributions to the field, this has been a curious one to wrap my brain around.

Ethics boards and opinion pieces alike tend to only focus on student use. Don’t get me wrong—AI use by students is quite high. In the UK’s largest annual survey, the share of students using generative AI for assessments jumped from 53% to 88% in a single year, and by 2025 the Higher Education Policy Institute found near-universal adoption. But I am seeing it at every level of academia, and for a plethora of different reasons and in a multitude of different ways.

• Academic journals, with drastically accelerated outputs and lower vetting standards, report being inundated with a flood of AI written or assisted papers
• Kobak et al., 2024 reviewed biomedical abstracts from 2024 and noted that one in eight showed signs of AI editing, while Liang et al., 2024 reported that up to 17% of peer reviews at conferences were machine-written, beyond spell-checking or minor writing updates.
• A survey from Frontiers enlightened to the fact that of roughly 1,600 academics across 111 countries, more than half had used AI while peer reviewing manuscripts, frequently against explicit journal guidance. On top of that, when the reviews/feedback were LLM-generated text, they were correlated with lower confidence in the work, were submitted closer to the deadline, and were then less likely to respond to author rebuttals.

And this is only the literature I can find by academics about academia!

As conversations about academic plagiarism blew up this summer after the public digital lynching of Professor Jason Arday, the fact remains that this industry and the media ecosystem is just as complacent in AI use as the tech bros. And this AI use is, on a technical, semantic level, academic plagiarism. While Professor Arday was not accused of using AI—and the accusations against him were a racist pile-on and fed by a predatory media machine, there’s a nonsensical line of thought here that has been getting to me. As one of my colleagues wrote on the matter:

“It was deemed preferable by institutions of ‘higher learning’ to fend off accusations of plagiarism, than to exercise the moral or ethical responsibility not to allow a colleague to be harmed. In so doing, UK academia rushed to defend its institutional reputation and academic prestige at the cost of ethical responsibility and human(e) conduct…

Pretending to care about the sanctity of academic standards…[they] fetishised scholarly rigour, proferring the pretense that the sector is not actually characterised by hastily conceived and rashly executed research due to the pressures of publishing as a precondition for getting or keeping an academic job.

As if academics don’t routinely use AI tools to research and publish, essentially plagiarising by lazily and illegitimately relying on whatever the AI algorithm hoovers up. As if academics don’t publish for the CV, the job promotion, or the REF, with the rush to climb the academic ladder overriding intellectual curiosity or the pursuit of knowledge at an unhurried pace. As if reputable academic publishers don’t use AI to produce and translate entire books to enrich their catalogue.”

On a broader scale, it must be noted that the utilization of AI (whether it is encouraged or tempted) is not an individual choice that I am looking to shame. Rather, we have found ourselves in a world where almost no industry is safe from the push for optimization, for more, and for more to be much quicker. The faster things can be done, the more competitive the stakes actually become.

We can (on a technical level) write more and publish more with it. When academic jobs are becoming increasingly precarious, and job security heavily relies on publishing output and the number of grants you can bring into the institution, it is no wonder that some choose to throw out 10 hastily written/partly generated articles a year rather than take the time to properly write one or two. I have seen this affect colleagues across disciplines who are required to use LLMs to keep up with the pace of competition.

I’ve also seen this phenomenon play out in what is called “slop grenades”, which was first coined by Tobi Lütke, the CEO of Shopify, referring to the increase of AI-generated work that is unedited and unchecked before sent along. It looks like polished, accomplished work but then it “explodes” into a mess that requires additional work from other teams to review and correct it.

It is like Jean Baudrillard said: “We live in a world where there is more and more information, and less and less meaning.” He was talking about mass media and the myth of expanding knowledge, and how it stages the appearance of communication and truth while actually exhausting and devouring its own content. To me, this is a complimentary descriptor of the impacts of AI-generated writing.

Right now, it feels like an automated rat race, and one we do not know how to get off. More than anything else, convenience takes precedence over depth. Critical thinking is no match for artificial generation when all that matters is speed and volume.

what is this doing to us??

You can (and should) go on your own rabbit hole about how AI writing is everywhere and how to detect it. Whether it is politicians to use it to write their speeches, or spiritual leaders using it to write sermons, it feels impossible to escape and the case-uses seem to be getting stranger.

The proliferation of this style of writing is beginning to make aspects of daily communication feel detached from humanity, a sham of a message. When we encounter communication without human idiosyncrasies it makes us question the creator. What are their principles, where does their passion come from, and in what context does it live in? If the ‘writer’ does not believe in their own writing, in themselves, then why should we believe in the message? It is an input and an output, rather than an intention and an action.

A friend of mine recently went to a wedding where he noted how a large percentage of the best-man speeches sounded lifeless, and uber-polished in a way that unusual for how he knew they spoke, and also all used similar metaphors for love and vague + broad phrasing. He was lamenting this to me, saying: “I am afraid now that one day when I die, someone I love is going to generate my eulogy.” It reminded me of a poem I have been returning to a lot recently titled “Please Use AI” by Jo Atkins-Potts that goes like this (and I recommend you read the full thing)

“Please use AI
when your mother dies.

Ask it what grief feels like.
Let it explain bereavement
in twelve clear bullet points.

…

Please use AI
the next time you fall in love.

Ask it how to know
whether someone is the one.

It will give you a faster answer
than watching a person become familiar,
learning how they take their tea,
which floorboards will wake them,
how their silence sounds
when something is wrong.

…

Why struggle
for the right words
when a machine can give you
beautiful ones?

Words untouched by shaking hands.
Words that have never sat beside
a hospital bed.
Words that have never known
the terrible privilege
of having something to lose.”

I guess what I am trying to get at—outside of the normalization of it in our jobs—is that the more we outsource our expression of love and frustration and shame and grief to AI, the more we create the conditions where only perfectionism is legible and desirable.

And yet, our brains know something is wrong here. We know from psychological studies that human brains actually show less beta wave activity (a strong marker of cognitive processing) when exposed to AI-synthesized voice rather than natural human speech. We know that just as smartphones made us stop memorizing phone numbers, AI assistants can cause us to stop practice-retrieving facts, leading to weaker biological memory pathways. We know that when AI summarizes a book or an article for you, your brain skips out on the deep processing stage. The hippocampus and temporal lobes need time and effort to move information into long-term memory. When you read a shortcut summary, the information usually fades quickly because your brain didn’t do the heavy lifting to store it.

Psychologist Betsy Sparrow identified a similar occurrence that was nicknamed the “Google Effect”. Her experiments showed that when people expect information to remain accessible online, they remember where to find it rather than the information itself. The participants also forgot information that they were ultimately confident they could look up later, but retained details they believed were not available online.

On one hand, I want to be cautious here: since the advent of writing, humans have always offloaded thoughts into external devices, like writing down notes. This has enabled us to have more complex lives and focus on things that really matter to us. When calculators were invented, mathematicians feared they would cause “laziness”; the invention of the camera made artists fear that painting and drawing would disappear because machines could capture reality in a second. So, this isn’t a totally new phenomenon. And it negates the ways that skills do evolve and focuses shift. At the same time, digital devices and AI increase our offloading at a level well beyond what the basic calculator, the analogue camera, and certainly pen and paper can do, and we are being primed to trust the information given to us from it without question.

I was reading research done by Professor Oliver Hardt, who studies the neurobiology of memory and forgetting at McGill University in Montreal. Hardt was sharing how his work has elucidated how prolonged use of GPS, for example, will likely reduce grey matter density in the hippocampus. The more convenient we make our technology—for everything— the more we will hinder our neuroplasticity.

There has to be a middle ground here. Wendy Suzuki explained on the Huberman Lab neuroscience podcast, “If we can’t remember what we’ve done, the information we’ve learned and the events of our lives, it changes us… [The part of the brain which remembers] really defines our personal histories. It defines who we are.” This is the point I’ve been marinating on the most. Memory stores our past and reflecting on it helps us process our experiences into insight and grow from our choices.

Because, we know that genuine empathy requires mentally + emotionally stepping into someone else’s shoes and reflecting on their own context, seeing them as 3-Dimensional beings just as we see ourselves. If we rely on AI to analyze people’s feelings or predict their reactions, we stop practicing intuitive human perception.

I saw a Facebook post about this a couple weeks ago that said it simpler: “We cannot read the room anymore, because we’ve stopped being able to read books.” And while that is a wee bit Boomerish, I’m not sure I disagree with them.

Another way of saying this is, when AI is both the shortcut and the main entity responsible for our emotional reflection, we become estranged from the richness of our own lives (Currie et al., 2025). It is alienation from the social world at a new level. Our personhood is no longer autonomous and grounded in our experiences, emotions, and thoughts about those experiences and emotions, but instead assembled across interfaces, platforms, and predictive logics.

We are seeing this even more as “emotionally intelligent AI solutions” are being peddled onto the market such as AI companions and AI therapy. This, along with the ability to ask a bot what a message means or if someone “likes us” or how to write a better breakup text, is not just the loss of introspection, but a larger change in how individuals relate to emotions at all. Outsourcing emotional intelligence to machines diminishes our emotional awareness and makes it difficult to negotiate subtlety in emotions without the help of the machines.

Social psychologists like George Herbert Mead and Erving Goffman (< 3) showed us that social skills are not inherent but learnt from trial and error, from observation, and from risk and correction. In the age of AI, I fear we have forgotten where we learned these skills. They are learned in the school of hard knocks: from being embarrassed after you said the wrong thing to someone you fancied in year 7, or accidentally being rude after you received a gift from your auntie age 14, to a blow up with a best friend in your 20s. We learn through conflict and tension and then, from thinking about why things might have gone badly. Avoiding and outsourcing this messiness means we stop learning why we do the things we do.

It is not just that AI writing is all around us, but that the successful market push, and ensuing inescapability of it means it is being used for emotional navigation and moral decision-making.

As a sociologist, I am far more interested in why people use it and how it is being used on a societal scale. I have no interest or time for shaming folks for using something that is promoted as a tool and pushed on them by employers and algorithms alike. I think a larger set of questions I am interested in around it is:

• what tasks/actions do people have difficulty with that they want help with?
• what scenarios do people feel scared of saying the wrong thing or unsure and awkward and therefore lean on AI to “fix it”, “clean it up”?
• what does this say about our reliance and capacity for vulnerability at large and how we teach/nurture those skills in the first place?

While not leaning into a witch trial of who is and isn’t using AI, it is important to make the point, over and over that it is atrophying our ability, both on individual and societal levels, to reflect. It atrophies our ability to see ourselves and others as real, complex, and human. It is destroying our capacity to know what to do about all our messy humanness.

why should we write and read at all?

A while back, I read this study published in the journal Judgment and Decision Making, where researchers had 1,682 adults read one of six short stories, three of which were written by humans and three by ChatGPT. The participants who read an AI-generated story rated it as more absorbing and of better quality than those who read a story written by a human. The New York Times conducted a similar blind taste test, where 86,000 readers took the interactive quiz comparing human and AI writing snippets. Over 54% of participating readers preferred AI-generated writing over human-authored passages.

I will admit, taking the NYT test, I chose many AI-generated lines over human ones, and when that was revealed, it put me in a bit of a tizzy. Yet, it actually makes sense. AI models are optimized via human feedback to appeal to broad preferences. It was an important reminder that all LLMs are trained on stolen literature (and blogs!) and labour that humans have constructed. It could not come up with a story about grief naturally because it does not have the ability to be sentient and to love and to know what it is to lose.

But, the more we get used to text becoming generated this way (cleanly and the best possible combination of Kurt Vonnegut + James Baldwin) the more we get used to a flattening of human experience and how we can describe it; because machines output statistical pattern matching, not lived reality.

It is not just AI that is being pushed on us, but the myth of optimization at large. The more we use machines to outsource our labour, our thinking processes, our emotional regulation, and our time, the more we are being primed to be machine-like. Our humanity requires consciousness; it requires failure and doubt and hesitation.

I’ve been thinking a lot about why humans develop and share stories at all. Why the need to put something into narrative account, to make people laugh, to recount a moment, to sneakily share an opinion or moral guidance, is all done via storytelling, and has been since our very start.

With writing systems, this became even more important. Again, it should be said that any new technology typically experiences pushback while being implemented. Socrates for example, argued that writing would produce forgetfulness and societal memory-loss of ideas/innovations because people would stop practicing oral recitation. This of course, was only remembered and now referenced because it was written down, in Plato’s Phaedrus.

Writing allowed thought to survive beyond mortality. It did not just preserve civilizations, but it also helped externalize consciousness. It gave humans the ability to make our inner worlds visible enough to examine them.

It is like how people turn to journaling, even if that is something they ~never do, during times of emotional intensity. For me, I often do not fully know what I think until I attempt to write it out. Writing isn’t efficient! It does not help us optimize our time, or hit our emotional KPIs with ease, but it does allow feeling to take durable form.

The act of putting pen to paper is so deeply, wonderfully tactile—you feel the vibration of the pen tip moving across the page, you create a distinct soundscape, you craft a visual personality. The act of writing calms an overactive mind and gives us space to process complex emotions. Whether it is angry scribbles and scratches or pressing harder on the page during a stressful thought, writing helps us empty out our brain well enough to be able to investigate what we are thinking and feeling in the first place.

Going back to Ollie, he said something in response to all my ramblings here that I want to share:

“AI can pull from swathes of training data to give me a statistically aggregated answer to solve my problems. But it can never hear me, listen to the feeling in my voice and understand how I feel. Sometimes we don’t need solutions! We need shoulders to cry on, ears to listen to us, and a place to put it so we can stop carrying it.”

last thoughts, for now

Writing this piece down has given me more ideas that I want to explore. It feels too reductive and almost like scolding for my conclusion here to be just to go pick up a book or to start journaling. AI adoption and decreasing literacy is not an individual set of choices or failures. We live in a society and that society has systems that are imposed upon us that we must navigate.

I will argue for, however, that we cling to our humanity through one another, and through stories, and through slower, more deliberate thought. Even if that is just going for a walk, or staring out the bus window on our commute home, or calling our mamas, or speaking to a neighbor in the stairwell, or developing a relationship with the staff at our local offie, or asking our pals questions about what they believe and feel about a thing, and just overall being okay with uncertainty and unknowables.

But of course, everyone getting a library card and using those spaces for learning about things you do not get to do in your 9-5, or reading a really interesting book, or reciting a poem to your friend in the stacks, or journaling on your lunch break— that too I encourage and will being doing right alongside you.