Welcome back to “thinking out loud with Sara.” Today, and most days, I’m thinking about biases within AI-generated content.
One of my summer projects was to create some materials to support faculty in their use of Canva with students. As part of that, I wanted to explore some of the new generative AI tools that Canva has introduced.
Before I started exploring, I heard a story on There Are No Girls on the Internet about Canva’s text-to-image tool flagging the prompt “black woman with bantu knots” as possibly resulting in unsafe or offensive content. This article from People of Color in Tech covers the story in more detail – and I highly recommend reading it.
Since I’m already a day late with this post, I’m just going to post some images from my initial searches, and give you the same prompts I hope to give students:
What do you see? What does it make you think? What do you wonder?
All images below are from prompts on July 26, 2023
These are all concerning (but not that surprising) in different ways. The search that really surprised me though was this one:
What? I reached out to Canva support about this, but was unable to get past canned responses to my questions and concerns.
As I started writing this post, I decided to try again, and see if Canva had addressed this. And I actually got results!
Then I decided to push my luck…
My response to this is probably not appropriate for a professional blog.
There’s a lot here to discuss with faculty and students, obviously, and there is a part of me that’s grateful to have such clear examples of bias in generative AI to use in conversations. But we all know that bias is not always this obvious – and is easily missed if we’re not consciously looking for it. How do we equip ourselves and our students to be on the lookout for these things? How do we craft prompts that account for these possibilities? How do we put the brakes on the rush to using generative AI while acknowledging that it is going to play a significant role in our lives? I don’t have good answers, but I know I need to keep asking these questions.
What are you wondering about when it comes to bias in generative AI? What questions are you asking? What questions are your students and faculty asking?
This post is going to be of the “thinking out loud” variety – I’d love to hear what you think and how you talk about these things with students.
I used ProCon.org early on in my career but never really loved it. If memory serves, a lot of their references were internal links, which was not particularly helpful when trying to find additional sources. I also found that the idea of pro/con was a false binary for many of the topics students were looking into. I hadn’t really looked at it closely in several years.
However, as I think about search instruction and the challenges of generating search terms that help students find multiple viewpoints I’ve found myself wondering about ProCon as a site that would help students get a high-level understanding of the arguments on different sides of an issue, and perhaps find language and key vocabulary that would help them find more sources. As I started looking I also noticed that the articles/footnotes provide information about sources outside of ProCon – but no direct links which is kind of annoying.
I decided to use some research I was doing for an upcoming class on banned books as a way to get a sense of how I might use ProCon with students (or for myself). And I am a little unsettled by what I discovered.
The Banned Books page starts with an overview of the issue, and then presents the top 3 “pro” and “con” arguments. I know that many of these arguments are based on personal beliefs and world view, but Pro 2 seemed to be pointing to some actual research and I wanted to follow that trail. To the footnotes!
And… hmmm. These sources are not exactly what I expected to find. But I did my due diligence and looked up the sources. I started with footnote 19 because despite the loaded language in the title I had the impression that it was citing actual research, and I wanted to find it.
Aha! A link to an actual study!
A study about pornography use. Which is not exactly how it was framed in either the article or on ProCon’s page.
Footnote 17, which references the American Academy of Pediatrics, actually points to an opinion piece on a site called Politichicks. This is the paragraph that references the AAP – but there’s no link to a specific source.
I did not do an exhaustive search of AAP recommendations, but I couldn’t find anything they’d written in favor of restricting access to books. A 2009 report I found actually recommended reading as an alternative to heavy media use.
There is a lot of nuance to the question of whether or not exposure to violent media causes aggression and I’m not prepared to unpack all of that. But suffice it to say it’s more complex than the author presents it in the Politicks article – and MUCH more complex than the sentence presented on ProCon.
I did not do a deep dive into every issue presented on ProCon, but I found similar issues in many of the pages I did explore. On the page about homework I found a dead link to an article from Monster.com – I don’t know what the article said, but I do know that if a student came to me with an article from a job search site I would have many questions for them about if this was the best source for them to use. On the page about corporal punishment there is a reference that points to an article that has since been deleted and replaced with a counterargument. I know maintaining up-to-date links is a challenge, but this one is particularly egregious (I’ve contacted them to make them aware).
I was hoping to find some information about how ProCon selects and evaluates sources, but despite being listed in the table of contents in the FAQs, the section on Sources does not actually exist.
So, what do I do with all this? The thing is, the blurbs on ProCon DO accurately represent the viewpoints and arguments on different sides of an issue. But what responsibility does a site that has as its mission “To promote civility, critical thinking, education, and informed citizenship” have to accurately contextualize those arguments and the sources being presented? They are presenting arguments on debatable topics (though the “debatability” of some of their topics is, well, debatable) and arguments like this are inherently biased. But being an informed critical thinker also means backing up your arguments with credible sources and accurately contextualizing the information you’re citing. There is something about this that feels like “choosing your own facts” and that is not what critical thinking looks like.
I think there is a way to use ProCon to build critical thinking and analysis skills, though probably not in the way they intended. I’m wondering about asking students to do the same sort of tracing sources I did and using that to evaluate the strengths of different arguments. I still think it could be useful for understanding what arguments people are making on different sides of an issue, but I’m wary of the possibility that students will accept the arguments at face value – after all, they’re citing their sources, and many students have internalized the idea that a website that cites its sources is a reliable source. But what if those citations lead to less-than-reliable sources? But what if those sources are accurate representations of the arguments being made? As you can tell, I’m going in circles on this.
I will be teaching an independent research class this winter, and I’m considering using this as an exercise with my students – how do we contextualize our sources, how do we evaluate arguments and rhetorical strategies, how do we keep an eye out for logical fallacies. Curious to hear how other folks think about and use resources like this.
This year I’m lucky to be teaching an Independent Research class – but I’m unlucky in that the first week of the class was the same week as the AISL conference, and also the last week of classes before Spring Break. In case you were wondering, it is not easy to launch a course when you’re across the country from your students. I wanted to get to know my students as researchers and get a sense of how they approach the research process; the challenge of figuring out how to do that when I wasn’t there definitely pushed my creativity and, if I do say so myself, I’m quite pleased with some of the activities I came up.
My favorite, however, was a task I called “What Does Research Look Like?” which I adapted from this assignment I found on Project Cora. The task itself is pretty straightforward, and my students had access to a wealth of materials in our design space – and those two factors combined resulted in some delightful AND insightful models of the research process. I was so delighted by what they came up with that I wanted to share a few examples. I probably wouldn’t have done this project had I actually been there for the first week of class, but it’s a project I’ll definitely do again!
Note: I asked students if they were okay with me sharing their work, and if so how/if they’d like to be identified
With this model we wanted to explore how research has a messy starting point with information coming from all directions and all different forms. The overall structure we wanted to highlight was an hourglass shape because research has a refining process as well as an application process. After the mess of ribbon at the top we glued on different colored and shaped feathers to represent various sources. Then, through the middle of our hourglass we glued googly eyes to represent the filtration process and analysis portion. At the bottom are bunches of pom poms which symbolize the pieces of info we synthesized at the end. Between the feathers and the pom poms we attached a green string to represent to process and connection.
Our model is meant to represent the different important steps in research: Initial idea; research through finding information and vetting sources; taking notes, and organizing important information; creating a final culmination of the information; and sharing the final product with others. We made each step a different level to signify the messiness of the process. Specifically how the process will take the researcher all over, through different sources, ideas, and questions. Though when looked at from the side all of these steps are in a line. This is to demonstrate a method to madness. The research process is crazy and unpredictable but in the end, it all comes together and creates a gain of knowledge.
When discussing what research looks like to us, we considered the often long journey that it can be. This long journey, is usually not linear, with many different ups and downs. When looking four sources your information may take you in new directions or even in loopdy-loops. When learning new things about your topic it may only inspire you to continue down the road. So, the thing that came to mind for us was a roller coaster. Our roller coaster has parts that are flat to represent starting points, it has a steep curve to prove a switch in directions, and it even has some swirly obstacles to show that you may encounter a few bumps down your road of research. Research to us can sometimes be scary or intimidating, but also very fun so we think our model is the perfect visual to how we feel about the process.
We view the research process as a series of increasingly specific gifts that you get to open to finally access the true knowledge of what you are searching for. To model this, we created a series of nesting boxes with bows on them which show how each source is narrowing in on what the essential question is. Inside of the final box, there is a brain that represents the final piece of knowledge. We thought of nesting dolls because during many of our research processes, we utilize information that we have gained from previous sources to seek out subsequent ones. If we had more time, we would have included things inside of the larger boxes or found a way to make it less linear.
It’s currently Spring Break at my school, so even though I have some ideas brewing from the AISL conference and the research seminar I’m teaching this spring, I wanted to share something a little cozier for this post.
As many of you likely know, February 1st was World Read Aloud Day. I also work with students in grades 6-12, which are not prime read-aloud years. But they should be! There has been plenty of research on the benefits of reading aloud, but for me the biggest benefit is the joy of sharing a story together.
I’ve only tried to do something once before with World Read Aloud Day – it was very elaborate and a middling success. I tried to do something that lasted multiple periods across the day, involving all kinds of teachers and while it engendered some positive feelings about the library, the reality was that it was poorly attended (even with the lure of hot chocolate!) and did not connect with students the way I had hoped it would.
But iteration is the name of the game, right? For this go-round, I decided to keep things much, much simpler. One storytime, during a time when most students are free, two picture books, and, of course, some cookies. I’m lucky to have a student ambassador for the library who I can consult with as I plan these things – she helped me pick the day and time, and also helped spread the word and generate interest.
I was expecting maybe five or six students to show (which I would have counted as a success!) but I ended up having almost 30 kids come to storytime! There were students from grades 9 through 12, and a handful of middle schoolers who were brought by the leader of their after-school activity. One student asked if they could respond to the story as I was reading it, and I said yes – which sparked a steady stream of “oh no!” and audible groans as I read about the trials and travails of poor Alexander.
This simple little storytime proved so popular that students requested it become a regular event. I did another storytime in February (and had a second planned that was canceled due to a snow day), this time inviting the faculty advisors to the Black Student Union and Black/African heritage affinity group to read aloud some of their favorite books. I’m hoping to continue to connect with other affinity group advisors for future storytimes.
I know I am often guilty of trying to go “too big” when planning, well, anything so it was both lovely and humbling to see so many people excited about such a simple program. And, thanks to Cindy Wray and Margaret Rhoades and their AISL presentation on creating a culture kids love, I now have a million more ideas for programs (big and small!) that will help me to connect to even more students.
I frequently tell students that using the same search terms over and over again will mean they find the same information and perspectives over and over again. But in my experience, students really struggle with how to develop a range of search terms. Inspired by a post of Tasha’s I wanted to try another way to help students think more expansively about what search terms they could use.
The class I worked with on this is doing research about repatriating culturally significant objects. They’ll be learning what they can about the history of a specific object, and then making an argument about whether or not it should be repatriated. This is a research task in which finding multiple perspectives is really important – and varying search terms is going to help students find those perspectives.
I started by talking about the difference between the words in your question and the words in your answer, using an example that a student came up with. This is a concept that many students find difficult to wrap their heads around, but this example really seems to help. We talk about how [impact] is not a term specific to their answer, but the different kinds of impact sun exposure can have are useful search terms – and also how sunburn/skin damage describe different impacts than vitamin D/seasonal affective disorder.
Next, I showed students how I might approach this task. I pulled passages from a few articles they’d already read, and highlighted terms that I might use in searching; I pulled out expert vocabulary, phrases, and the names of organizations and legislation. I then gave students two articles about the repatriation of an Alutiiq kayak that was held by Harvard’s Peabody Museum. One article was from the Harvard Crimson, and the other was a press release from the Alutiiq Museum. Working in small groups, students highlighted terms and phrases that they thought could be useful in their search.
The list of terms they found was amazing! This list below is in addition to the ones I pulled out from the passages I read. This activity also allowed us to correct some misunderstandings about what might make for effective search terms.
As students shared terms, I asked them to note which article they had found the term in. We had a brief discussion about how the terms differed between sources; next time I do this I need to devote more time to this part of the lesson as it’s a valuable part of understanding how different terms help find different perspectives.
After this lesson students have a bank of search terms to return to as they search – and, hopefully, a better understanding of how to find effective search terms.
My goal for this blog post was to have some organized thoughts about ChatGPT to share, but I think the best I can do is still just some disorganized thoughts. There is, of course, lots to talk about and think about, but I’ve been spending some time thinking specifically about the role ChatGPT could play in research. I’m going to spare myself from trying to write transitions and just go for some bullet points.
Students can struggle with finding an appropriate source to build background knowledge on a topic. I experimented with asking ChatGPT to give me a paragraph about different topics students are researching, and the writing it produced was full of expert vocabulary, important ideas, and potential search terms. It could be useful for modeling how to use background sources, but also for helping students find a jumping-off point when they’re new to a topic.
This is not an idea I came up with, but I’ve had fun playing with it: ask ChatGPT to write you the table of contents for a book about something. I was working with some colleagues on a course about media influences, so I asked ChatGPT to give me the table of contents for a textbook we could use. It gave a really solid outline of what we could think about. When using it for research, it could provide some guidance about what subtopics you could explore. You can also ask ChatGPT to expand on different chapters of your imaginary textbook!
I’ve played a little with asking ChatGPT directly for search terms, and am still deciding what I think about it. Admittedly I’ve given it pretty vague prompts, so the search terms have also been pretty broad. I did notice, however, that it generated search terms that represented different political viewpoints – and it also encouraged me to be more specific in my research. 🙂
I think prompt crafting is going to become an important skill. When I gave ChatGPT a vague prompt, I got unimpressive answers. As I refined my request, the responses got better. The advantage of ChatGPT is that I can keep asking for refinements to the previous response. This means that I need to clarify my own thinking so I can ask for what I want – either on the first try, or by evaluating the response and making further requests. Being clear on what you’re looking for (both for yourself, and when creating a search) is such an important skill and the conversational nature of ChatGPT could provide some practice.
I’m aware of the ethical and practical concerns around ChatGPT and AI (and my colleagues can assure you that I will share them at even the hint of an opportunity), but I’m also aware that our students will have access to these tools as they move through the world. I’m hoping we can skip the years of hand-wringing (*coughcough* Wikipedia *coughcough*) and instead help shape the conversation about how we can meaningfully and ethically make use of these tools.
So, how are you thinking about the role ChatGPT can play in your work?
When I got the calendar alert reminding me I had a blog post due soon, I thought I’d write about ChatGPT – it’s been a big topic of conversation at my school and I have been in many really interesting conversations about what it is, what it means, and how we can use it. But then I caught Covid (my first time!) and the post-break return to school is always hectic so I haven’t really had the brain space to put my thoughts into words.
Instead, I thought I’d share about an activity I did with our Media and Its Influences class yesterday. It was our first day of classes back from break, and a few of us are doing a “guest teacher” unit for the month of January. We wanted to both get students’ brains back in gear, and also learn a bit more about how they think about and evaluate news coverage.
I went to Newseum’s collection of front pages from key moments in history and picked four events – Hurricane Katrina, the Supreme Court decision about same-sex marriage, the 2016 election, and the release of the Mueller report. I chose four front pages from each event, looking for a range of front pages – in tone, layout, location, etc. – and printed them on big pieces of paper and posted them around the room.
After introducing the activity, we set students loose to examine the front pages and add notes/reflections responding to the following prompts:
What are the “vibes” of what you’re seeing?
What stands out to you about the word choice?
The physical layout of the article?
What do you notice or wonder about regarding the different ways the event was covered by these sources?
What questions do you have?
After everyone had a chance to look at all collections, we divided students into groups and gave them each a collection to discuss and to share takeaways with the group.
It definitely took some prodding to get students to offer the “why” of their interpretations, but with some gentle nudges they had some really great insights. One student noticed that almost all of the pictures from the same-sex marriage front pages were of white women (which was not something I’d done on purpose, but when I went back to look at the rest of the front pages was accurate overall). Another student noted that one of the front pages about Trump’s election looked like a poster, so we talked about the role of “front pages” historically. After a student noted that he’d never actually seen the front page of a newspaper (because no one in his house subscribes to a print paper) another student countered that “home pages” for websites work much the same way. There were lots of other great observations and discussions as well, and there will be lots we can refer back to and build on in order to deepen their understanding of how news media works.
One of the recurring tasks on my weekly to do list is to find a middle school student who wants to recommend a book – and it’s often one of the highlights of my week. Every week during Middle School meeting a student gets up to talk about a book they love and that they think their peers will enjoy as well.
The format is pretty simple. Students offer a couple of sentences about why they like the book, and who they would recommend it to. But it’s the conversation it takes to get to those sentences that I really love. I usually stop by a classroom looking for volunteers, and on good days I’m greeted by eager students who want to share a book. Sometimes it takes a touch more cajoling, but once I get one-on-one with a student they light up talking about the book they’re recommending.
After we turn our conversation into a more concise format, we make a slide with the book cover and recommendation, and the student gets up during the meeting to share about the book. It’s a great opportunity for students to get some low-stakes public speaking practice too!
How about you? How do you have students share book recommendations?
I’m sure many of you saw the headlines about TikTok being “the new search engine” for Gen Z. And if your colleagues are anything like mine, they wanted to talk to you about what it meant, how search works on TikTok, and if I would be changing any of my teaching strategies as a result. I had had several of these conversations before it suddenly occurred to me – was this even true? How was this determined? Why were we all so ready to believe this?
So I did what I tell students to do – I applied some SIFT strategies to this information. The first step, of course, being to Stop and pay attention to what my reaction to this story was – a reaction which can probably best be summed up by grandpa Simpson.
Okay, I’ve accounted for my biases. I saw this story in the New York Times and other publications I’m familiar with, so I didn’t spend much time Investigating the source. I did, however, have an inkling that my emotional reaction to this information might have been shared by some of the people (likely similar in age to me) who were reporting on this story.
My efforts to Find trusted coverage were also pretty brief. This story had been in a lot of places – though I did notice that a lot of the stories relied on anecdotes. Hmmm…
Finally, I decided to Trace these claims to the original source. The source seems to be an interview with Google Senior Vice President Prabhakar Raghavan in which he said:
“In our studies, something like almost 40% of young people, when they’re looking for a place for lunch, they don’t go to Google Maps or Search,” he continued. “They go to TikTok or Instagram.”
Which is interesting! But is also very much not what was typically being reported. The story I linked to also goes on to note that the data isn’t public – another red flag.
Luckily enough, I was recently invited into a Chemistry class to do a lesson on evaluating popular science reporting and applying SIFT strategies to all those articles about what “a recent study found.” I was able to use this experience as an example, and talk to students about how my own biases and assumptions got in the way of my critical thinking. I love being able to talk with students about how I use the strategies I teach in my own day-to-day life.
The SIFT strategies are tremendously useful and I teach them all the time, but I’m realizing I often give short shrift to the Stop part of the process. For the lesson I did with the Chemistry class I asked student to specifically note the reaction they had to the story before doing any of the other steps. Looking at student reflections was fascinating, and gave me lots of insight into how students react to the sources they encounter – I will not, however, be turning these anecdotes into a story for the New York Times. 🙂
There is a rhythm to a school year. The energy is far different in September than it is in the weeks before spring break. It’s something I love about working in schools – you can really feel the “seasons” of the school year and the shared sense of community that comes as we move through the year.
This year, however, I am out of sync. I started the school year on bereavement leave, and shortly after returning had a bad fall that kept me out of school for a few additional days. This week is really my first week back full-time; it’s my start of the year, but everyone else is already in the thick of it.
I always knew I relied on that time before classes start to get ready for the year, tie up loose ends on summer projects, and set up my space ready for the year. It is so disorienting to be trying (and, I’ll be honest, mostly failing) to do those things once the year is underway. I am trying to get ready for a year that is already underway. My space still feels disorganized, I know I missed information from opening meetings, and I am still introducing myself to new colleagues. Colleagues are eager to jump into collaborations, but I am still getting my bearings.
I am also not coming in with the type of energy and focus I typically have at the start of the school year. Grief has made my brain… fuzzy, and I’m struggling to remember names, ideas, and tasks, or to do the types of lesson planning I usually do. It is frustrating to feel out of sync in so many ways.
I know there has been a lot of conversation over the past several years about the need to take care of ourselves and to be realistic about what we can and can’t do. Those ideas have gone from theoretical to very, very practical for me, and I’m trying to be clear with myself and others about what my limitations are at the moment. Yes, I’d love to talk ideas but I need more time to process. Yes, I can recommend a book but it will take me a while. Yes, I’m happy to meet with students but I really need a heads-up about what you all are working on. Yes, I want to be a part of things but everything is still too much for me.
I’m grateful for colleagues who have been understanding and offered support (and even encouraged me to take more time off). It is a gift to be a part of this community. And I know that people are accommodating in part because they know how difficult the last couple of months have been for me. But I also wonder what it would be like to do this all the time, not just when we are in acute need. As much as there is a rhythm to the school year, all our lives also have their own rhythm. What would change about how we work together if we were in tune with the cadence of each other’s lives?