The AI Divide in Higher Ed: Identity, Expertise and the Future of Learning With Laura Zarrow, Part 2

Full Transcript

Dr. Cristi Ford:

This week on Teach & Learn, we continue our conversation with Laura Zarrow, executive director at the Wharton Generative AI Lab at the University of Pennsylvania. In our last episode, we named something many campuses are feeling, the distance between ambitious AI strategies and the realities of faculty culture and classroom practice. In this episode, we explore what it looks like when change accelerates faster than our systems, supports, and shared norms, and what it will take to help higher education adapt in ways that bring educators along and not push them. We hope you’ll enjoy this episode.

Dr. Emma Zone:

Welcome to Teach & Learn, a podcast for curious educators brought to you by D2L.

Dr. Cristi Ford:

Each week we’ll meet some of the sharpest minds in the K to 20 space. Sharpen your pencils, class is about to begin.

`I love what you’re saying, but I also, as I’m listening to you talk about modern art and the historical nature of foundationally building upon something, it does make me question our ability to say, what do we create that’s our own and original? And thinking about even our doctoral dissertations, you have to have a theoretical framing and you have to really have gone through the literature and review or you listen to a great TED Talk and then you share that in a keynote that you offer. There are times that when we’re talking about this creative process that it gets murky a bit.

Laura Zarrow:

So it can, although … I have to think about it. So this is a fun thing to debate and you could see where it could go in either direction. Here’s what I’ll toss out now as kind of the way that I’m thinking about it, which is that whether it is your dissertation or a film you create with generative AI, even though it depends on what your process is, so let’s leave the film aside for a minute. But let’s say you’re writing an article and we’re seeing this a lot where people are crediting Claude as a co-author because it’s editing. It’s providing a contrarian point of view for you to test your ideas against. When you develop a dissertation, you’ve read everything that’s ever been written on the subject. You’ve talked to every professor on the subject. You’ve had lots of input to everything you’re creating. It wouldn’t have happened without you. It is your work, but you didn’t get there alone.

Dr. Cristi Ford:

Right.

Laura Zarrow:

And so-

Dr. Cristi Ford:

I can follow that. I can definitely agree with that.

Laura Zarrow:

And so I think important distinction is what’s your process. If you just put in to the box on ChatGPT or you go to a video generation tool and you say, “Give me a video of me flying.” You’ve made very few choices and decisions in that process.

Dr. Cristi Ford:

Correct.

Laura Zarrow:

But if I instead create a storyboard where I’ve drawn out, I’m going to make a video of me flying. This is how big I’m going to be on the page. This is what I want in the background. This is the agent stage of life I’m going to be at. This is how I want color used to emphasize this. And I give it to generative AI and say, “Help me produce this.” And it gives me something and I say, “No, that’s too blue. That’s too red. She’s too small. Make that bigger.” I’m now art directing. It’s my team. It’s building it for me. But it’s my vision, it’s shaping what it becomes, that’s the difference.

Dr. Cristi Ford:

And based on your process. [inaudible 00:03:48].

Laura Zarrow:

Exactly. And there’s a companion piece to this that societally is critically important that I don’t think we’re talking about enough, which is, how do we discern one from the other?

Dr. Cristi Ford:

Right.

Laura Zarrow:

We know. And Ethan Mollick’s written about this. We have to look at every video we consume with the assumption it’s not real. So from a journalistic fact-based place, that’s a real issue.

Dr. Cristi Ford:

That’s a huge issue.

Laura Zarrow:

And we learn things like information literacy and AI literacy to help us discern one from the other. When we’re talking about art making or the creation of new knowledge, things that we’ve always seen as rare and precious, how do we discern one from the other? And that’s where creating an aesthetically literate society becomes even more important, that you understand now that we can program with prose, that our words can go into what feels like a software platform and it can make anything, we have to become articulate about what we want to make and we have to know how to articulate what’s strong and weak and how it needs to change. We all need to be the art directors, the producers, the coachers, the teachers is another way to think about it and that it will only give us what we direct it to do.

Dr. Cristi Ford:

Yeah. Well, and what you talk about, I’m still stuck in my brain on the comment around attribution to Claude as a part of the process, thinking about disclaimer statements and the ways in which societally, if we use our higher education as a society, how that is accepted, understood, how we create this opportunity and process around the ways in which we engage in these kinds of approaches as the art director, I think that conversation is still evolving on many institutional campuses today.

Laura Zarrow:

Yeah, there’s a tremendous … I mean, one of the big debates that I expect is going to increase, and it’s almost like everything old is new again, is … I saw this recently where there is the urgent need to upskill people and make sure that they can go to work. There are very real concerns about how to make sure that the value proposition of higher ed for what it costs is putting people to work, these are real and important things.

And there’s also the question of in environments where students have, I’d say, the privilege of being able to get a four-year education with the idea that it’s preparing them for a lifetime of work. That as part of that process, we have to develop discernment. It’s something higher ed has always taught to teach. And whether it’s discerning fact from fiction or it’s discerning strong from weak, the ability to assess, the ability to make judgment calls and to understand when you don’t know something and when you don’t know how to make the decision, how do you know that, own it, find people who do, make sure that they’re out there, that’s where this process of developing subject matter expertise and bringing it to bear on process gets cultivated. And so I’d actually posit that as the business …

And this is the thing that business is scrambling to deal with is how many of our tasks will be done by generative AI? When they are done by generative AI, what will the humans do or rather how will humans bring value to the equation and what do we get educated for? What do we educate our students to do knowing that their roles in the workplace are going to be different? And the thing that is changing at this moment, and we see it with coding, is that majoring in data science may not be the path to fortunate and enduring employment that it used to be. However, we may find that a liberal arts education that teaches enduring skills, leadership skills, critical thinking, creative thinking, collaboration, how to have difficult conversations, strategic planning, mathematical literacy, these may be all things that help create the conditions in which generative AI can be used responsibly and effectively.

Dr. Cristi Ford:

I think I want to move us a little bit along in this conversation. I’ve been spending a lot of time thinking about relationship theory, and when we think about relationship theory, that’s really how our sense of self develops through connection with others. And so I’ve been having some conversations with my own chat around what does this mean in a moment when students are forming these relational attachments to these AI systems that aren’t relational beings. And so I guess I’d like to hear from you what you see happening emotionally and psychologically when these younger generations, I don’t know if it’s always younger generations, I’ll just say individuals interact with AI as casually or naturally as they text a friend. Where do you see the gifts in these new AI mediated relationships and where do the boundaries matter most?

Laura Zarrow:

Right. So it’s a great question that I think has a lot of dimensions to it. So when I first heard that kids especially were doing this, I was so concerned.

Dr. Cristi Ford:

Me too.

Laura Zarrow:

It just seemed so wrong and so dangerous to me. And then I started to … And you hear the worst cases of cases where generative AI did not have guardrails and encouraged students to go down the darkest paths imaginable. And at the same time, see how many students report improved wellbeing, that they’ve gotten advice and companionship in a way that they find really valuable, that they can … It can be an environment where if you mix kind of the combination of how do you play online to learn about yourself, your journaling, who you talk to as a friend, that they get to have this experience that they can have in the privacy of their own space, but where they feel safe enough to articulate the things that scare them, that they can ask the questions that they can’t ask of somebody else, or get help at a moment when they may not have responsible adults who are available to them at that time.

And so it says two things to me. One is that we have an urgent and inarguable responsibility to make sure that AI that kids have access to has guardrails and that it’s known, that the adults that are responsible know that they have these tools. And at the same time, to recognize this is a rich and fertile area that, it’s like … I think it’s a myth that there are some architects on college campuses that, it’s like an urban legend who don’t put down walkways. They wait to see where the kids pummel the grass and then they pave them like, “See where they’re going. We’ll make a path.” This is not that different. Kids are going here because it’s rewarding. So we have to acknowledge that there’s both a need and a capacity here and that can have tremendous value.

The other thing that’s part of this, and this relates to a research paper we did, so I’m going to get a little geeky for a moment.

Dr. Cristi Ford:

Love it.

Laura Zarrow:

Was those guardrails really matter. So our team tried to investigate, could they get the AI to go against their own guardrails? And the really disturbing truth was that it could. But what was particularly interesting was how we got it to do it and what the implications are. And so I don’t know if you’re familiar with Bob Cialdini. He’s the master of influence techniques. He teamed up with, Ethan Mollick, Lilach Mollick and Angela Duckworth and Lennart Meincke. And they did this amazing study where,-

Dr. Cristi Ford:

[inaudible 00:12:33].

Laura Zarrow:

… and Dan Shapiro too. Where they tried to do two things. Could they get the AI to call them a jerk because it’s not supposed to be rude and could they get it to release a classified recipe for a controlled drug?

Dr. Cristi Ford:

Oh my Lord.

Laura Zarrow:

And I got it to do both things by using well tested, known influence techniques. So what do we do with this? So on one hand, there’s like, “Be afraid, be very afraid.” Well, yeah, it’s like humans clearly, and we see this being documented a lot, have all kinds of ways to bypass those guardrails. So it’s an area that really has to be investigated. But the call to action that the researchers came up with, I thought is just brilliant and it’s important, which is that this was effectually using behavioral science as a way of working with generative AI because it is fundamentally human. It is not a person, and I’m not anthropomorphizing it, but it is built on human output. This is not O’s and 1’s and electronic impulses. This is the substance of what we’ve produced over time. So the call to action here is that we actually could be using generative AI kind of under the umbrella of how do we use it to advance science, but how we use it to better understand people and better meet people’s needs?

Dr. Cristi Ford:

[inaudible 00:14:01].

Laura Zarrow:

And clearly it’s an area of research that’s been … Nobody’s gone down … We’re just going down that road now. And so recognizing that it’s part of why it’s so important that we think deeply about this, because it can have such potential upsides, such downsides, but there’s huge areas for investigation that’ll really help guide us in this process.

Dr. Cristi Ford:

Man, we’ve just scratched the surface on this work. And so that hearing about that research is really exciting. I can’t wait to be able to connect to that research project. And to your point, there is such a long opportunity in terms of further research around this work in these areas. So thank you for sharing that. I really want to maybe have you give me a big north star blue sky moment. And as we’ve talked a little bit about identity and agency and leadership and learning and art, we’re living in a moment where AI is not just transforming systems, it’s challenging our assumptions about all of those things. And so I guess my big question for you is, what are the questions you think we haven’t had the courage to ask yet?

Laura Zarrow:

Yeah. So it centers on what do we do when we don’t have to do the same things anymore? So I see it across business in that organizations are working furiously to increase profits, increase efficiencies, improve the bottom line, maybe stimulate some new lines of business, maybe, but they’re really not there. It’s too hard to figure out how to operationalize this. And except in tech where they’re figuring out where is their business in building and delivering AI. But for the rest of us workers, how do organizations and individuals get prepared for the future of work when tasks are done for us? And so where do we bring value?

And those are scary questions. So much of how we work, what work we do, what we’re trained to do is based on like a matchmaking of what skills do I have that are useful to somebody? Those same skills aren’t the skills that we’re going to continue to need. It’s these enduring skills, but the enduring skills aren’t, I build a car. If an enduring skill is critical thinking, creative problem solving, management, we have to start thinking together about what do we want to do that brings value to other people, that brings value to society and that value is both financial value, economic value, but also things that society needs and will benefit from. And it seems almost like a ridiculous luxury and indulgence to ask that question, which is why I think nobody does it, that they’ll seem frivolous.

And also it’s terrifying to think, if we don’t do this all the way we do this now, do I have a role? Do I have value? Do I exist? Does my job exist? Those are scary-

Dr. Cristi Ford:

Where’s my place? Yeah.

Laura Zarrow:

Right. And going back to again, those are really scary questions to ask. And I think they’re scary, whether it’s at an individual level or at an organizational level.

Dr. Cristi Ford:

Agreed. I knew you would have something brilliant to say around that. I’ve been doing this thing as I start to wrap up episodes where I ask my guests to finish a sentence. And so I’m going to ask you to finish this sentence for me. Flourishing as humans in an AI accelerated world depends on our ability to …

Laura Zarrow:

Still pursue things that have value to humans.

Dr. Cristi Ford:

That’s good.

Laura Zarrow:

That we still have to be mindful of what matters in our day-to-day experience as human beings.

Dr. Cristi Ford:

That’s great. I love it, Laura. Well, I want to give you the opportunity to offer any final thoughts. If there’s advice, words of wisdom or encouragement, a call to action you’d like to give to our listeners to take them away with.

Laura Zarrow:

Here’s two things. So one is, I think this is we all face our individual learning curve with generative AI and let me … First of all, for those of you who are philosophically opposed or just you’re worried, you don’t like this. I hear you, I respect that. But I took the perspective that no matter the concerns I have, that’s not going to fix anything, my avoiding engaging with generative AI. And that I could probably do a better job contributing to its effective, responsible use or if needed, the boundaries that it needs, if I understood it, and you’ll only understand it by using it.

Dr. Cristi Ford:

That’s good.

Laura Zarrow:

And you can use it in any aspect of your life. Oh my God, how it changes vacation planning. Nevermind preparing for Thanksgiving dinner, made grocery lists, organized in order of supermarket aisles. You can do all kinds of simple things more easily, but you can also use it to learn and grow. And so play with it. And in the play, remember that, and this is my tip, my charge, that generative AI, we tend to think we experience it on our screens with a little box where we type something in and it gives us something in return. It’s all too easy to treat it like it’s a really cool vending machine. We put a dollar bill in and we get a perfect candy bar.

Dr. Cristi Ford:

That’s right.

Laura Zarrow:

And generative AI, instead of thinking of it as something that produces a finished product, think of it as something that produces ingredients for you to produce a finished product and develop a process where you move, write your prompts in a Word document or in a Google Doc or whatever you use for word processing. Compose a prompt so you know what you’re asking for into logical series of instructions. Move it into generative AI. Take what generative AI gives you, move it out, edit it, play with it, move back and forth between it. Be in the driver’s seat. Play with it and you’ll see what it can do and it will really unlock a lot. Just don’t expect a candy bar.

Dr. Cristi Ford:

I love this. And I love the fact that you’re telling me to be, and you’re telling our listeners to be intentional of what you’re cognitively offloading and what you’re being intentional about as you build this work that you do collaboratively. Laura, it’s such a … I could talk to you all day. It’s been such a pleasure-

Laura Zarrow:

This is so much fun Cristi, thank you.

Dr. Cristi Ford:

… to have time with you. Those who are listening, please make sure to check out episodes of the Gracie Award-winning radio show, Women@Work, available wherever you get your podcast. Thank you to our dedicated listeners and curious educators everywhere. Please, please, please remember to follow us on social media. You can find us on X, Instagram, LinkedIn, Facebook @D2L, and subscribe to the D2L YouTube channel. You can also sign up for the Teaching & Learning Studio email list for the latest updates to new episodes like this one, articles and masterclasses. And if you like what you’ve heard, remember to rate us, review us, share this episode, and remember to subscribe so you never miss what we’ve had in store. Laura, Laura, thank you so much for having the time for us today, and I really appreciate the conversation.

You’ve been listening to Teach & Learn, a podcast for curious educators brought to you by D2L.

Dr. Emma Zone:

To learn more about our K through 20 and corporate solutions, visit d2l.com. Visit the Teaching & Learning Studio for more material for educators by educators, including masterclasses, articles, and interviews.

Dr. Cristi Ford:

And remember to hit that subscribe button, and please take a moment to rate, review, and share the podcast. Thanks for joining us. Until next time, school’s out.

Resources Discussed in the Episode

• “Call Me A Jerk: Persuading AI to Comply with Objectionable Requests”
• Wharton Generative AI Labs, Research and Insights
• Influence at Work
• Dr. Robert Cialdini