How I Stopped Worrying and Learned to Love the Bot

In Spring 2023, Anth160 was my first class to meet in person, so they became my ChatGPT guinea pigs.  I generally bring up a series of very simple questions about health on the first day of class in order to get students talking and thinking about what they already know and what anthropology can bring to the study of health, illness and healing.  The three questions I queried ChatGPT3 about in January 2023 were ‘What is health?’; ‘What does it mean to be healthy?’; and ‘What should I do to stay healthy?’ I posted the answers in my Powerpoint, but before introducing the whole notion of ChatGPT, I first asked the students for their input into these questions. 

Because this is largely a medical anthropology course, there tend to be a lot of nursing and health science students in this class. This means they generally enter with a solid background in health issues, but most do not have any background in anthropology when they start my course.  My job is to draw on their health knowledge and to help weave it into the anthropological studies that explore the issues from a more social, symbolic and political-economic perspective.  Throughout the semester, we learn about health practices and ideas from around the globe and explore our own health practices and ideologies here in the US.  These simple introductory questions help illustrate some of the ‘givens’ in our society and provide an opportunity to explore ways our thinking about health is influenced by culture and the environment. The ChatGPT responses to this question added a new layer for us to talk about. 

As with most years, many students remarked that health is an absence of illness or disease.  Some brought up mental health.  Several suggested that health and healthiness equated having energy and participating in sports or exercise.  Healthy living included ‘eating right’ and avoiding junk food or too much sugar.  Drinking enough water, doing yoga and meditation were mentioned as ways of staying healthy.  Usually, I use this discussion as an opportunity to get students to dig deeper, provide details to explore specific health practices that they used growing up, which often helps to highlight the cultural and social aspects of their health ideologies. 

This time, I let ChatGPT interrupt our discussion and used its answers as a jumping board for both a critique of ChatGPT technology and an opportunity to elaborate on their own health ideas and practices. 

Below are the answers provided by ChatGPT3 to my 3 questions:

 

What does it mean to be healthy?

Being healthy generally means having a balance of physical, mental, and emotional well-being. It encompasses a range of factors such as eating a balanced diet, getting enough exercise and sleep, managing stress, and avoiding unhealthy behaviors such as smoking or excessive alcohol consumption. Additionally, having access to proper medical care and maintaining a positive outlook can also be important aspects of overall health.

What is health?

Health can be defined as a state of complete physical, mental, and social well-being, and not just the absence of disease or infirmity. Health is a dynamic and ever-changing state, and includes a wide range of factors such as genetics, lifestyle, environment, and access to healthcare. Health encompasses a range of physical, mental, and emotional aspects of well-being and is essential for overall quality of life.

What should I do to stay healthy?

Here are some general guidelines for maintaining good health: -Eat a balanced diet that includes a variety of fruits, vegetables, whole grains, lean proteins, and healthy fats. -Get regular exercise, aiming for at least 30 minutes of moderate activity most days of the week. -Maintain a healthy weight. -Get enough sleep, aiming for 7-9 hours per night. -Manage stress through techniques such as meditation, yoga, or deep breathing exercises. -Avoid smoking, excessive alcohol consumption, and drug use. -Practice safe sex and use protection to prevent sexually transmitted infections. -Wash your hands frequently to prevent the spread of germs. -Stay up to date on recommended vaccinations and screenings. -See a doctor for regular check-ups and screenings. -Practice good oral hygiene by brushing and flossing your teeth twice a day.

Note that these are general guidelines, you should consult with your doctor or health professional for personalized recommendations based on your individual health needs and lifestyle.

 

I also added in the bonus question “Are Doughnuts Healthy?”  The response is telling, falling squarely into the ChatGPT3 model of always balancing conflicting viewpoints.  ChatGPT3 seems driven to provide at least two sides to every issue, even with regards to the healthiness of doughnuts! 

 

Doughnuts are typically high in calories, sugar, and unhealthy fats such as trans fats, which can contribute to weight gain, high blood sugar, high blood pressure, and other health problems. They are also typically low in important nutrients like fiber, vitamins, and minerals.

However, it's important to note that doughnuts can be made in different ways, and some doughnuts may be healthier than others. For example, doughnuts made with whole wheat flour and natural sweeteners, such as honey, may have less sugar and more fiber than traditional doughnuts. Also, if you opt for a smaller size doughnut or share one with someone, it would decrease the amount of calories and sugar you consume. Therefore, doughnuts should be consumed in moderation as part of a balanced diet and healthy lifestyle, not as a regular part of one's diet. If you want to enjoy a doughnut, it's better to choose a healthier option and balance it with other healthy food choices.

As you can see, much like the ChatGPT newsletter, these responses are perfectly adequate.  Most people would likely agree with the information provided in each answer, even the suggestion that eating smaller doughnuts is healthier than eating larger ones!  In fact, one student immediately raised her hand and noted that the ChatGPT answers were almost word for word copies from her health textbook’s definitions. Now this could be interpreted as a huge red flag for the plagiarism capabilities of ChatGPT.  Students can easily produce a textbook answer using ChatGPT, if that is what is assigned to them.  I would venture to guess, however, that most of us do not want the textbook answer from our students.  We want students to explore the ideas in their own words.  We want them to demonstrate that they understand what they’ve read, which means we want them to explain the concepts and bring in examples, and, at least in anthropology, I want my students to draw on their own life experiences. 

One important lesson here is that ChatGPT does not provide the types of responses that we are looking for, even for a question as simple as ‘what is health?’  This is a good thing.  We can build on this.

During the class discussion, when I asked students about the language of the ChatGPT responses, everyone agreed that it did not sound like a reflection they would write about health even though it mentioned many of the exact same ideas they came up with in our initial brainstorming activity.  In fact, most agreed that it did not even sound like it was written by a human.  They critiqued the language as dry, boring, abstract, and lifeless. Where was the lived experience of health in these answers? And what did the ChatGPT really mean by the generalities that it put forward about health?

This gets to our second learning moment. ChatGPT was my chance to show them what I do and do not want to see in their informal reflections about the readings. The ChatGPT answers provided us with a foundational text that we could build on.  This serves two important goals.  It sidesteps the tendency of students to repeat the textbook answer in their writings, and it allows them to explore each part of the response in more detail and to fill the writing with life, which is exactly what I want them to do. 

In our discussion we picked out a few concepts from the ChatGPT answers and started digging deeper and personalizing the questions. What do you consider a healthy diet? What are some actual foods that you eat? Why do you consider them healthy? How do they make you feel when you eat them?  What do you actually do in your day to day life to stay healthy? Who do you go to for advice about diet and exercise? Who are your role models? This line of questioning helped the students bring the answers to life.  They personalized their answers with examples from their lived experiences.  They spoke in their own voices and they clarified through these explanations how they understood the concepts and ideas.  What the students realized through this exercise is that ChatGPT does not know us and our lives.  It cannot replicate our experiences and our thoughts or emotions.  It can formulate a hypothetical experience, but it does not know us personally and cannot draw from our personal repertoire of life events (at least not yet!). 

This was a fundamentally calming realization for all of us.  Students could see the types of answers that I look for and realized that their writing was actually preferable to the abstract textbook language produced by ChatGPT.  Their experiences matter and their voices matter.

For me the in-class discussion about the ChatGPT was a relief because the inadequacies of the bot became clear (it was easily recognizable as not written by students) and because of the learning opportunity the bot afforded by helping students move past the textbook language and toward explaining and developing their answers more thoroughly and comprehensively. 

Does this mean students will not turn to ChatGPT to plagiarize? Of course not.  Does this guarantee that they will read the material that I assign them? Also no.  But they are engaging with the ideas and hopefully enough of them will realize that they do not want to sound like chatbots.  Certainly, when they read the attempts of ChatGPT to speak in slang or more like a specific person (as we’ll see below), it becomes blatantly clear how inadequate the bot still is, even the updated version of ChatGPT 3.5. 

Building on the experiences from my Anth 160 class, I developed similar questions for my introductory course in anthropology.  This time, however, I added a new angle.  What happens if instead of asking a simple question, you ask ChatGPT to write ‘in the style of’ someone, say, perhaps a college student? Would these results be equally as sterile? Would they follow the pattern of offering ‘both sides’ arguments? Would it still be easy to identify such responses as products of AI technology? I was particularly interested in this question following the introduction of the updated ChatGPT3.5, which has allegedly become more sophisticated and nuanced in its responses.  I needn’t have worried.  As you can see from the texts below, the bot did level up its style and creativity, but it by no means is able to accurately replicate the style of language I requested. Let’s start with the introductory question, ‘What does it mean to be human?

1. What does it mean to be human?

"Being human refers to the characteristics and qualities that define the species Homo sapiens. These include the ability to think, feel, and communicate complex ideas and emotions, as well as the capacity for self-awareness and consciousness.

Other characteristics often associated with being human include the ability to use tools, create and appreciate art, and engage in complex social interactions. Additionally, humans possess unique physical characteristics such as bipedalism, opposable thumbs, and a large brain-to-body ratio.  The concept of what it means to be human can also encompass moral, ethical, and spiritual dimensions, and can be understood differently by different cultures and individuals."

 

With this straightforward question I went through the same steps as I had with my Anth 160 students above, and the results were similar.  One student this semester commented that the text seems to have been written by a sociopath.  Psychiatric diagnoses aside, this did segue nicely into a discussion of how we could humanize the response and make it our own. 

I then brought out several ‘in the style of’ essays for them to read and critique in small groups.  I’ve only included one of the styles here, but the results for all of them were quite funny.  ChatGPT3.5 wrote in the style of a neanderthal, a Valley girl, a chimpanzee (after several attempts which it initially refused due to the inability of chimpanzees to write or speak English!), a generic teenager, and the essay below, which is allegedly in the style of an 18 year-old man from Queens, NY.  One of my students was kind and brave enough to read the text out loud for the whole class.  As you read through it, you can imagine the giggles this text elicited from our students, who completely tore the text to shreds.  It incorporates some modern slang, but it reads as if it were written by an adult who has almost no experience listening to or interacting with actual teenagers in NYC, and this holds for all of the texts that I put through the app.  The ChatGPT3.5 answers read like parody or, in the case of the teen scripts, straight from Steve Buscemi’s skit “How do you do, fellow kids,” where he tries to blend in with the high school kids despite looking every bit of 55.

 

ChatGPT3.5: Write an essay by an 18 year-old man from Queens, NY about anthropology.

Yo, what's good? So, check it – we're talking about anthropology, and honestly, it's kinda wild how it's all about diggin' into our past, present, and future. I'm an 18-year-old dude from Queens, NY, and I gotta admit, at first, this whole anthropology thing sounded like some fancy-pants stuff. But guess what? It's way cooler than I thought.

So, like, anthropology is this whole deal where we're peepin' into what makes us humans tick, from back in the day to right now. It's like having a time machine to check out how our ancestors did their thing – what they ate, how they lived, and maybe even how they partied (you never know, right?). It's not just about bones and old stuff; it's about gettin' inside their heads and, you know, piecing together the story of our people.

But hold up, it's not just about history. There's this vibe called cultural anthropology that's all about how different groups of people live their lives and do their thing. Like, imagine you're chillin' in Queens, and then you jet off to some remote tribe in the Amazon rainforest. Cultural anthropologists are doin' that – hanging with different crews, checking out their styles, their slang, and, like, the way they vibe with each other. It's like being a fly on the wall in a hundred different spots at once.

And then you got linguistic anthropology – like, what's the deal with the way we talk? I mean, I'm from Queens, where every block's got its own lingo. But these folks are straight-up detectives, figuring out why we say "deadass" or "fuhgeddaboudit." It's like a puzzle, and they're piecing together how we connect through words and, you know, the little nods we give each other when we're on the same wave.

Now, archaeology ain't just some dusty dudes in hats – it's like a real-life treasure hunt. These archaeologists are like the modern-day Indiana Joneses, but maybe with less whip action. They're digging up old stuff – pottery, tools, maybe even a gnarly weapon or two – to understand how people lived before smartphones and Netflix. It's like unlocking secrets from the past, and that's pretty dang rad.

But here's what's straight-up awesome about anthropology – it's all about, like, respectin' the differences and realizing we're all part of the same human fam. Like, Queens is this melting pot, you know? And anthropology's like that on a global scale. It's about gettin' rid of stereotypes and really getting what makes each group unique and, at the same time, how we're all connected.

In the end, anthropology's this crazy journey that's like, way more relatable than I thought. If you're into discovering why we do what we do, how we got here, and what's next, then anthropology's the real deal. It's mind-blowing, eye-opening, and it's all about being a part of something bigger. So, yeah, Queens got love for anthropology, no doubt.

 

I hope it goes without saying that this does not sound like it was written by an 18 year-old from Queens, NY.  What it does do, however, is help calm my nerves.  ChatGPT is not as human as some would have us believe, and we should be able to identify it if our students decide to resort to it for their writing assignments. 

As concluding remarks, I’d like to talk about a philosophical shift I’ve made this semester in response to ChatGPT and the writing endeavor more generally as well as some of the results I’ve witnessed in my students’ writing.

For the past several weeks I’ve been participating in a writing bootcamp. My goal is to finish some of the writing projects I’ve been working on for some time (including this newsletter - goal achieved!). As I’ve been going through the lessons, I’ve realized that much of what I’ve been learning is also applicable to our students.  One of the primary issues we’ve been discussing is procrastination. Why do so many people procrastinate, even when they enjoy writing? Why do so many people start projects that they do not finish? A helpful sourcebook we have drawn from is Rosanne Bane’s Around the Writer’s Block: Using Brain Science to Solve Writer’s Resistance. A key point she makes in her book is that resistance to writing is normal. There is a physiological explanation. Our brain is simply reacting to a perceived danger - it’s a fight or flight reaction to negative experiences we have around writing, particularly experiences of being judged. The good news is that the brain can be re-trained and new habits for writing can be developed, but it takes time, and it involves a shift in focus from final product to process.

What does this mean for our students and how is it connected with the problems of ChatGPT? While plagiarism is an issue that is largely out of our control, we can take steps to help students enjoy writing more, which in turn will lessen the temptation, at least for some students, to turn to such tools as ChatGPT. Why do students turn to such tools in the first place? Laziness is not the only or even primary reason for many students. Cheating is something found among both high achievers and struggling students. High achievers can fall to the pressures of getting straight As. This can lead some to seek outside help because they fear that anything less than an A is a tragedy. This focus on grades can turn the process of learning into a nightmare.

Other students have such negative views of their own writing that they have convinced themselves that they cannot write; that they are not writers. Often, especially at the beginning levels, students are not comfortable enough with their own understandings of the material to write about what they’ve read without restating the original text.  Some may even think that is what we want to see in their work. They may replicate the texts in the readings because they want it to sound good, and they fear they will not be able to reproduce something as good as the textbook or ChatGPT. 

Shifting students’ attention to the process of writing and learning rather than the final products and grades can help alleviate some of their anxieties and encourage them to explore their own writing journeys. This can take many different forms. This semester I’ve been talking with my students more about the writing process and ways of building writing habits.  We’ve talked about taking small steps. No one would expect a person to be able to run 10 miles if they haven’t run in years.  We would expect them to start small.  That’s what we do with writing. Writing takes practice.  Changing habits is difficult. Our brain freezes and thinks it’s protecting us.  We need to take baby steps and reward the brain for small successes in order teach it that writing is not the same as a wild beast and is not going to kill us.

One thing I’ve introduced this semester is an Excel tracking sheet that I adapted from Monica Hay’s writing bootcamp, and I’ve shared it with my students.  Each time we write - we can track that. Then we get a reward for our progress and process and not just the end-product. I’ve also shifted my thinking about grades. They do get grades for their writing, but I try not to make it punitive. I want them to be able to explore and practice. In my in-person class, this means adding more free writing and short writing sprints.  For formal writing there are chances for multiple drafts (which is a standard practice for WI courses), and as long as they are fulfilling the basic requirements, there should be nothing for them to fear in their writing.

In the processes I’m trying to make a shift in my entire teaching philosophy - away from the focus and fear around plagiarism and ChatGPT taking over the world - and toward process - and the small interventions that make writing easier, more practical and more enjoyable for my students (and myself). We’re only three weeks into the semester, and I’m already seeing good results. 

After discussing the AI issues with ChatGPT in class and re-designing my questions to personalize their discussions more, I’ve been extremely pleased with the first sets of responses students have written this semester. Their Discussion Board responses touched on different aspects of the readings that resonated with them. They brought up experiences from their own lives that related to the topics in the articles I assigned. I’ve been happily surprised by the depth of engagement with the material they have expressed in their answers. Different topics have resonated with each student. Some were interested in the breadth of topics covered in anthropology. Some connected with the effects of colonialism and discussed how colonial notions about skin color had affected their own self esteem. Others were drawn to the descriptions of life in other parts of the world. And one with a more philosophical bent wrote about an extra article I posted on ontology and stones. They were drawn to this different way of thinking about being in the world. It’s a complex article, but from conversations in our class it was clear that these were issues the student thought about, so I wasn’t entirely surprised by the response. 

Another writing ‘win’ that I’m thrilled about happened just recently in my Anth 101 class. On the first day of class, when I announced how excited I was that our class was WI, one of the students in the front row frowned and shook her head. I tried countering her reaction with something along the lines of “Writing is fun and allows us to express ourselves and build our own stories!” I was hoping my enthusiasm would be infectious and rub off on the students. I at least wanted to put a good spin on it on the first day, and I also offered them some techniques to help incorporate writing into their busy lives. “Do the bare minimum.”  This is a mantra from the bootcamp. Start small and start to build the writing habits. Five minutes a day at first. More frowns and sadness. The following week we did some in-class writing. This was for practice but also served as a way to help them get to know one another and to give them a jump start on their first formal writing assignment, which is a reflection on their identity. Students wrote for 10 minutes about something that is important to them, and then they shared their stories with each other in small groups. These discussions were so lively that I allowed them to continue for longer than I had planned. But what surprised me the most was what happened the following week. The very same student who had scowled at the thought of writing asked me to read a draft of her assignment that she had completed two weeks before it was due!

This is certainly a small sample of students. I’m not saying these ideas are a panacea and that my students never use ChatGPT. I also don’t know if this engagement with the material this semester followed directly from our discussions in class about ChatGPT and writing, but my students’ responses to the ChatGPT in class and their writing assignments so far this semester have helped me to stop worrying so much about the bot, and I hope these experiences can help ease some of your anxieties as well.