Kyle Bylin: Hello everyone and welcome to the Infinite Stream Podcast. This episode explores the current state of artificial intelligence in higher education and the radical future that powerful AI could create. Today on the show, I'm joined by Peter Shea, the Assistant Dean for AI integration at Middlesex Community College in Massachusetts. Peter has a background in English education and began teaching college writing back in 1993. He began looking at AI in 2017 as a way to give students more feedback on their writing. Today, his work involves helping an entire college think through how AI should fit into teaching, student support, and everyday operations. We're going to explore the plan Middlesex is currently developing, the messy middle of trying to make change happen, and what community colleges can do to prepare their students for an uncertain future. Peter, thank you so much for coming on the show today.
Peter Shea: Thank you, Carl. It's a pleasure to be here.
Kyle Bylin: Peter, to start out, could you tell me about your role at Middlesex? What does the assistant dean for AI integration actually do? And what are you ultimately hoping to accomplish across the college?
Peter Shea: I think my role in has three different functions. One is to help lowering the anxiety around AI as an as a enormous technological and cultural change. Two help faculty, staff and students figure out what connection, if any, AI plays with their legacy practices and approaches. And three, help them envision How AI can be used to do entirely new things that we could never do before thanks to its unique affordances.
Kyle Bylin: And can you speak on the anxiety side of things? Obviously the pace of AI continues every day, and over the past four years it's certainly disrupted higher education a lot of different ways. What kind of anxiety are you seeing around campus?
Peter Shea: Well, you know, my campus, my college has always had a place for innovation. We were the first community college in Massachusetts to have an online program. And we were exploring things like immersive learning in the virtual worlds over a decade ago. So culturally, we you know, we weren't as shocked as other colleges were. We even played with a pre-generative AI tool around twenty eighteen or twenty nineteen. but what I am seeing in the larger Higher education space where I also operate as an advisor. I would say year one of AI's arrival was very anxiety-producing and disloc, you know, dislocating. Year two and three were still anxiety, but less so and a cautious engagement with AI. And this now, in what we are, I guess, year three or four. I'm sensing still engagement. People understand it's it's part of the world and we need to forward, but there's also I think more hostility towards AI than I I sensed in the past. because they're seeing the public narrative around AI has been associated with job loss, it's been associated with environmental impact of data centers, it's been associated with potential apocalyptic scenarios. AI is good at many many things, but it's not good at getting good press. and higher higher education is one of those fields where it has been harder to put to get a consistently positive relationship with AI. And despite a lot of effort, there is still considerable ambivalence, more ambivalence than I think in than in other fields.
Kyle Bylin: What do you think that the challenges are of overcoming this kind of bad PR problem that AI has? For people paying attention, as you're saying, the headlines are talking about job loss. The next week they're talking about, you know, existential risk to humanity. And meanwhile, we're all kind of still talking about a chat bot, we're talking about agents, but it seems like the potential of these capabilities can get lost in the conversation when we make it specifically about the environment, specifically about data centers, copyright infringement, and like all of the other kind of challenges that led to the AI that we have today.
Peter Shea: definitely. I you know, I I think one of the disadvantages for AI is that it's it's public public representatives, you know, in terms of the heads of the AI companies, are not the kind of people you would necessarily want. They're not a PR person's dream, to put it this way. They're not the best spokesman. There's no there's no Steve Jobs among the AI crowd. There's no one of of equal charisma. And a few of them have what you might call negative charisma. so
Kyle Bylin: Mm-hmm.
Peter Shea: also often the messenger is the part of the message. So that's been a bit of an issue. but there is a big you know, after all this negative feedback, there is a big there is some degree of growing awareness that AI has an enormous potential to to do more good than bad. In fact, the New York Times, I think it was On oct it was October second. they published a guest study called Six Charts That Show Just How Much We Need AI. And it addressed such things as we've gone through a productivity slump in the past few decades. We went through a lot of growth, economic growth, and then we just kind of like stalled. And we need some sort of mechanism to to to to reboot our both our economy and our you know our general culture. And AI is the kind of technology that has the possibility of doing that. So I I see it as as a as a great tool to to restart the the economic engine that you know sort of stole the nineteen seventies for the United States. and I think as a tool properly used can can increase people's sense of agency rather than diminish it. unfortunately, right now, we're we're still People are still using it like a sort of supercharged Google, which it's not, you know. I think it was is it David Sow of MIT who just published a book a few days ago where he talked about the excitement around the personal computer was because, as Steve Jobs said, it was a bicycle for the mind. So that was the bicycle for the mind. He said, AI is a motorcycle for the Which is to say it can get you a lot further than a bicycle, but there are dangers with a motorcycle that you don't have with a bicycle. so the I think it's a pretty good metaphor. I think we can get to places that we've never been able to get before. and I I'm very excited about the possibility. But unfortunately, particularly in higher ed, which is a inherently conservative industry, we are intent on looking back and simply preserving things, even things that have not been working out very well, because we're Attached to tradition rather than looking forward to what we can build that we can never build before. And I I think a lot of you know, my public communication around AI is trying to get people to look away from the preservation and more towards the creative construction. Or I like to say, stop, stop talking to the firemen and start talking to the contractors, you know.
Kyle Bylin: I think that's such a a rich metaphor, this idea that Steve Jobs said that the computer is like a bicycle for the mind. And I love your extension of that metaphor, that it that it's more like a motorcycle. What I would add on that is this idea that we all brought our creativity, our intelligence, our imagination to computers. And for For the vast majority of us, the keyboard was the main way that we could to talk to an intelligent machine, right? But these days, through voice prompting and obviously machine intelligence, now we can basically communicate in the with the computer through English, but then have that transformed into code and any other modality that we could potentially create. So in many ways, We are now the bottleneck in the sense that the limit of our imagination in terms of what we can envision is really the only limit as to what is possible right now. And it's interesting to kind of communicate that at a time when we're really focused on: well, is the model smart? Does it get everything right? As opposed to the human side and the empowerment that these new technologies unlock.
Peter Shea: Yeah, we are we are new, you know, whenever a new major technology emerges, we spend at least a decade negotiating our relationship with the technology. it's a it's an awkward dance. but we've been here before. People forget that. If you're familiar with the history of technological integration, the you know, things which seem so seamlessly a part of modern life were often resisted or misunderstood in the first few decades. And A good example would be the automobile. we tend to imagine that the minute Henry Ford opened his factory, everyone was buying cars, and that's not the case at all. It took a long time to shift from a horse based travel methodology method to an automobile centric one. And in many in many key industries, resistance to changing from horses to automobiles or mechanized vehicles was was fairly strong, particularly in the military, which had a lot of people, senior officers who were invested in the cavalry and the horses, and who did not want to see their expertise in that area become obsolete. So and I I think you see that a lot with college educated people and public intellectuals arguing against AI. Part of the thing AI threatens is there is that The value they often provide is their expertise, but also the scarcity of their expertise. And when their expertise becomes less scarce, it becomes less valuable. And so there's a threat to their own economic well-being, which I think manifests itself often in s and maybe not consciously, but it manifests itself a lot in the public criticism of AI. but I again I think you know. You can rely upon AI overmuch and allow it to weaken your natural capabilities. But you don't have to. You have a chance to define your relationship with the technology. You're not passive, you have agency. That's what people forget. You know, some people are gonna choose to use it lazily, but the night the same people use all technologies lazily. You can say no to certain practices and say yes to others. And for me as an educator, I'm more interested in AI as a tool that can augment human ability and develop human and develop the ability that then lets later gets augmented. So I see it not just as an extension of human ability, but as a tool properly used, which can give us the kind of immersive learning experiences that enable us to get really, really good at things and which are full potential. And an example of that I would be is that. Is we simulation-based learning. It used to be very difficult to build a meaningful simulation, which was unfortunate because they were uniquely powerful in in producing really high quality learning for people, particularly in demanding fields. I mean, that's what got us to the moon. That's three astronauts in a in a space simulator. But now with with AI, we can we can build and scale up simulation in a way we never could before. I think that's right. I think there are two things about AI. in education that have been overlooked in all the fury. And one is AI has effectively closed the door on a lot of practices which were flawed, but we continue to do so because people were attached to them. It's like Balboa burning his boats behind him. It's going anywhere, like there's no there's there's no way back. You can't we we can't claw claw our way back to that world no matter how many blue books you take out and have students write in them. The second thing is AI has potential to build learning experiences and scale them in a way it's been utterly impossible throughout history. And that I think is revolutionary. And that I think is getting overlooked in all the public furor about AI.
Kyle Bylin: I'd love to hear more about that last point that you made, which is that these educational experiences at scale are now possible through the different tools and services that these AI companies have made available to us. Can you talk about what your thoughts on that are, or maybe some examples from Middlesex that you guys have done in trying to roll out educational resources?
Peter Shea: Yeah, sure. One of the things I like about AI is that you don't you can build things in-house. I mean, I've been involved with educational technology for thirty years. And that, you know, apart from basic web pages, that almost inevitably required c contracting with a vendor. At Middlesex, we've built a lot of things in-house using an AI API key to to power it. So if you view AI as a form of energy, then you build the things you wanted to energize and you then you can experiment with them. And we built a number of AI powered tools in-house at very little expense that we can implement. And I think that it creates a new it creates a new dynamic for educational technology because before you were you were entirely dependent on vendors. You know, w one of the most exciting parts about the early stage of the internet was back when. Ordinary people were creating their own web pages and deploying them. They weren't having to pay someone to do it. That was a really liberating experience. And that's what unleashed a lot of the excitement about it. You didn't have to pay someone. I think with AI, now we have the same experience all over again. We we we don't have to we don't have to go to a vendor, we don't have to go to a company organization. We can build on our own. And I think that's enormously liberating. and I it's not only that, it's also very cost effective, which I think is another factor which is going to really help you know increase the demand for a AI-based tools. the problem I think is in higher education, it's we're trying to attach a an engine to a horse. and then when it doesn't work, we blame the engine. It's not the engine, it's the we have to have conversation about the horse. In fact, if I ever wrote an article about it, I probably call it we need We need to talk about the horse. Because a lot of people in education see themselves they're not they don't see themselves in being tra in as a as a comparison. They don't see themselves as being in the transportation business. They see themselves in the horse business. And I say, no, we're in the transportation business. We're trying to get people from here to there. And the horse has been traditionally the way we get there. And I know we've been doing it for a long time, but now there's a better method. And that's really it requires a real paradigm shift for people, which is obviously always difficult for anyone.
Kyle Bylin: In a presentation that you shared with me, you described three parts of your current job, which is counselor, facilitator, and futurist. I'd be curious which of those takes up the most time right now in terms of the capacities you kind of have to fulfill on campus in those different roles.
Peter Shea: I think in the beginning, counselor was a big part of it. And I knew it would be. But I am spending more and more time in the futurist role, which is what I really want wanted to be to begin with. but you have to lay the ground. I often say in in a job like mine, the absolute you don't you don't need to have a technology degree or background, but you need to have like an anthropology or psychology background because it's not the challenges aren't technological. Someone's already figured that part out. It's the human change management element that is essential. And you can know a lot about AI, but if you don't know much about change management or the culture of your organization, you're going to fail. So
Kyle Bylin: I I can absolutely see that. to pick up where you kind of left off there, I definitely agree that that some aspect of this requires us to be a bit of a futurist. And in fact, futures thinking is what I have integrated into my own AI lectures to students, which is that I I paint this picture that the frontier AI labs since their beginning. Have been working on this idea of AGI or artificial general intelligence. And some of them are even saying that either they've achieved it now or they might be close within a few years. You combine that with different predictions from the World Economic Forum's future jobs report and how the market could shake out by 2030. And you have a world that In the four years that a student spends with us, could be very different by the time that they graduate. And so I try to empower them with prompts and tools and concepts to where they can talk to AI about their specific field and get personalized recommendations about what opportunities are emerging, what careers are potentially going away. and how to essentially use AI to keep abreast of all the different news and events in their field. That way, if they pay attention to this, if they track the changes that are happening, AI can be something that potentially empowers them rather than something that they feel is happening to them.
Peter Shea: yeah, I I definitely think one of the things that colleges have to do is figure out how to talk about AI with their students. They have to have the talk and they had they they can be candid about their concerns, particularly in regards to the way that the misuse of AI can inter intervene with productive this productive struggle for l needed for learning. That's perfectly legitimate and absolutely required. But the same time they have to acknowledge that their students are going out into a world where it's going to be shaped by this technology. And they have to give them some sense of how to begin developing their own relationship with it. And this is a great opportunity for them to help shape that relationship positively rather than leaving students up on their own. And when they don't do that or they're just simply throwing up their hands, then they're I think they're failing in their obligation to students.
Kyle Bylin: Absolutely. And like building on this discussion, we've talked about this idea of potentially becoming over-reliant on AI, potentially outsourcing both your doing and your thinking. You started in this space exploring how AI could be applied to student writing and how it could give them feedback. I'd be curious, like looking back, what did you see then? And now how do you think about AI's involvement in the writing process today?
Peter Shea: I think AI properly used by both the faculty member and the student can can enormously enhance a person's ability to develop their writing skills. But again, it's it's how you it's not whether you use the tool or you don't use the tool, it's how you use the tool. Like a scalpel. I mean a scalpel can can be very dangerous in the wrong hands. you can hurt yourself. Same with a power drill. If you use them properly, you can do remarkable things. Students would have to understand what part of the writing process, particularly when they're learning, should not be outsourced to the AI and which which the student themselves have to own. On the other hand, an AI's ability to analyze a particular student's text and give them targeted feedback. Is uniquely valuable. I've never seen anything like that before in my life. And I've seen AI give better feedback to an example of student running than than I've given because the AI has such unique analytical powers. And feedback is crucial to developing people as writers and communicators. So I I I see it as a largely positive relationship.
Kyle Bylin: It's been really interesting to watch the idea of writing evolve as the models have gained in their capacity. I remember sitting at the research desk next to another librarian about two years ago, and they talked about this idea that AI could could fabricate a paper essentially, and it could also come up with fabricated citations. And at the time, You know, this was seen as like a limit on the system, as if it would always be there. And as you could guess where this story is going, a few months later, deep research came out. The AIs got better at doing research, synthesizing information together, and producing citations. And since then, I believe it has got gotten even better and potentially even more accurate. To where, if you look at that study done by MIT, they suggest that most undergraduate homework at this point could potentially be automated. And I think that like raises a question in terms of what exactly is the homework that's being assigned to students? What are the outcomes that we're trying to achieve? And are these the same outcomes that they're ultimately going to be asked to create in the workforce, essentially?
Peter Shea: yeah, you know, again, the the idea of yeah, the idea of homework itself is is gonna be challenged, particularly for K through twelve. I mean if I had a K through twelve student, I I wouldn't I would not I wanna I would want some way to track my child's progress, but I wouldn't want I ho in my experience homework was often assigned as a way of reassuring parents, as a way of so they can see what their stu their kid is doing. rather than just relying upon you know report card grades. my view is that if you know if you're spending six hours a day, five days a week in school, there's really no reason to be taking work home with you. You I mean you should be able to you should be able to do all your productive learning work at school. And when you're not in school, you should that should be your time to spend. So I'm not a big I'm not really mourning the death of homework. and obviously it's different for college students. I just think we have to The other thing that we have to deal with is is is figuring out new ways to assess students' thinking and skill development that rather than our our legacy because the attachment to essays and written exams was it was perceived as the most practical way of of assessing students' knowledge growth. that's no longer the case. We need new assessment tools. And so that's going to be a big, very big conversation in higher education because you know, we like to say that we're in the education business, but we're really in the credentialing business. We we sell credentials. People can people can learn things without us, but they can't get a credential without us, at least not in most fields. But if our credentialing system is broken and everybody knows that our business model is is you know under threat. So I think you know, the question is really: how do we are there better ways to assess? what students are learning and and how they can apply it.
Kyle Bylin: I'd be curious what have been your takeaway so far. I, like you, went to the teaching and learning conference with AI in Florida. And it was interesting to see all of the different ways that people are incorporating chatbots and maybe early stages of agents into the classroom. But ultimately, the assessment of the work was still kind of a challenge. Because you were trying to figure out at what point do you assess the students' work and what kind of like work are you attempting to s to assess? So I'd be curious like if you learned anything from that conference in terms of innovative assessment and to what extent you guys have been able to apply it at middle sex in any of your different departments.
Peter Shea: Well I think you know In my presentation at the conference in Orlando, I did an analysis of all the proposals and I noticed that a lot of them were around building frameworks, which I thought was interesting. You know, frameworks is our version of of a of a of a dike. You know, we're trying to keep the water from coming in. But you you can only get the dike so high up. At some point, you have to realize that you gotta retreat and and and stand your ground somewhere else. at Middlesex, for example, we explore using simulation tools to capture assessment data so we can track people's decisions over time. It's much harder to cheat in in a simulation-based assessment than than in conventional ones, and it can be it it can render much more interesting data, I think. I you know, I've always been struck by the potential of learning analytics. but I don't think learning analytics ever achieved its full potential because there were no ways to really extract interesting data using our conventional methods. I mean I used to go to presentations where people would get excited to say, you know, we're tracking when students log into their LMS and we're tracking when they log off. And it looks like many of them are only looking at the assignment three hours before it's due. And I remember one group of presenters that were so pleased by that. And I'm thinking this is the best you can do in terms of the data you can sift. But if we build tools that track decisions and actions, we can get much more interesting data to to assess student learning than we ever did before. And that's the kind thing that excites me.
Kyle Bylin: Absolutely. Zooming out a little bit, I'd be curious to know how would you describe the way that middle sex has kind of responded to this AI moment? Where do you believe you guys are making progress versus what are the areas where you feel like you're still trying to play catch up?
Peter Shea: Well, I don't think there are too many places we have to worry ketchup because I don't think there are many peop people in the lead to begin with. we're still kinda all absolute beginners. I think that the fact that there's a full that I have a full time role doing this says a lot about where the college is at and the fact that we have a three year strategic plan. And we just recently acquired a a Title Three grant around AI integration to support that. So I think that helps establish, you know, where we are on the field.
Kyle Bylin: And when you kind of zoomed out to the Teaching with AI conference and you saw what everyone else is doing, what was your kind of analysis of the overall field? I certainly felt like there's a lot of pilots, there's a lot of interesting spaghetti being thrown against the wall, but no one really knows what is sticking. And because we have to deal with, you know, faculty and students in real time, it doesn't necessarily always guarantee buy-in. And one year you could be all in everyone's favor and have people attending your various events and stuff like that, but the next semester it could be a total ghost town and you can run into, you know, trouble sustaining that same momentum.
Peter Shea: That's why I think those of us who are at the frontier of trying to advance this work need to speak with one another on a regular basis and support one another as a community of practice because we all have the same issues. And the buy-in is going to fluctuate and it's gonna be different from institution to institution. I think a a small I think a thousand campfires is hard to sustain. I think there needs to be more coordination, more sharing of innovative practices that are that are replicated and can be scaled up. I think we need to have some sort of association where we're we're we're trying new things and a having a shared experimental space and when we come up with something that seems to work, share it out with the others so that we you we can all invest in one another's success.
Kyle Bylin: I'd be curious as you look across the landscape of higher education and the different AI initiatives that you've seen roll out, are there any that you felt were especially on point and that you wished you could, you know, bring to your own campus?
Peter Shea: I'm I'm very impressed by what they did down at Miami Dade College in terms of the creation of NIAC, the National AI Consortium, where they immediately recognized the need to help design courses around the AI fluencies that certain employers would want. I think they really took the bull by the horns and and you know, had a vision and an operational plan. And I'd like to see some of my initiatives flourish the way theirs have. I think they've been a really they've been a really great leader in this field.
Kyle Bylin: love to hear more about the initiatives that you're trying to do at Middlesex. What are the bright spots in terms of successes that you have or things underway this semester that you feel like are really going to help get things off the ground?
Peter Shea: I like using AI to tackle bottleneck problems. you know, one of the ways in which we're looking at it, having AI review writing samples and then giving the students advice, not not telling them what to do, but giving them advice as to what section of English they should probably enroll in when they come to the college. the volume of of of Writing samples that we see every semester is is is a lot and and you know, the process of evaluating them can be slowed down and AI can really help address that in a speedy way. That's one example. Another example would be a tool we had that takes a job description and then identifies potential questions the applicant might get during an interview. I think for our students that's that kind of is very valuable because many of them are not at the point where they can envision what they're gonna be asked in a job interview. So I think using AI to solve problems that have seemed unsolvable before, I think is really one of the best applications of it.
Kyle Bylin: I I definitely agree. I think this concept of the wicked problem is something that aligns itself well with AI, just in the sense that if we look out to our communities and the broader world, there's all types of wicked problems, whether it's like human poverty and food access to even the higher levels of climate change and management. And What we need more than anything is people who can use their domain knowledge, their unique taste and agency to leverage these tools to potentially find answers or solutions to some of these wicked problems that perhaps we may not be able to do on our own, or now suddenly small teams of students can make tremendous effort towards even in a semester, let alone during their entire undergraduate curriculum.
Peter Shea: you know, I I think I don't think most colleges are going to I don't think most colleges are going to do radical changes to the curriculum. I don't think they're equipped to do that. I think they have to be adaptive of it. And I think some colleges will become very fluid in adaptation and curricular change. One thing that's gotta change is curricular view and adoption has to move quicker than it has in the past because I I recall I recall talking with someone who once had the job titled Provost for Innovation and then he left that job and went back to being a professor. And I said, Well, why is a great title? He said, No, it's not. He said, We we we we would propose a course that we really saw a need for, and it would take three years to get it through curriculum review. So I think things like that. I the the the structure of of programs and disciplines may not change, but I think the curricular review process will change. So that's the one at least my my prediction.
Kyle Bylin: I could definitely see that as being a challenge. It's ones that I've ultimately seen in my own work too. I'd be curious, something that I've wondered about for the past couple of years is this idea that we had the divid digital divide, which talked about access to computers and the internet and all of the opportunities that the digital age could unlock. But now increasingly I think there is an AI divide emerging. Not just in colleges, but also in curriculums, in the sense that some institutions are providing full access to paid tiers, whereas others have students that are ultimately footing their own bill. But beyond that, you know, access to the intelligence and the paid tiers, I also think is like the cultivation of the mindset of what is possible. And what I mean by that is that there's just Something distinctly different from like learning what AI is, how a large language model functions, and basic prompting capabilities when compared with some of the education that is emerging on campuses like Stanford, for example, where I can imagine that vibe coding was established very early on. Students are prototyping with these tools, they're designing products and services. Rather than kind of being worried about this generation of information and how it can be used for plagiarism, they're talking about the use of code and how it can be used to build businesses that potentially further the lives of students and maybe create opportunities for them that that weren't even there before until AI democratized coding itself to everyone.
Peter Shea: Yeah. And what's more, in certain professions, the ability to work with AI is going to become essential. Now it's not going to be equal across the board, but I've heard from like the CEOs of IT companies that have said one of it bluntly to to me and a group of others, if your computer science students don't use AI, don't bother to send them to our job interview.
Kyle Bylin: Absolutely. And like the reality is, like, I feel like that is becoming more of the norm in a previous conversation that I had with Jessica Stansbury. she talked about this idea that the AI portfolio is increasingly becoming a thing that employers are looking for. And I thought that was really interesting because in the user experience design and research profession, portfolios have been a thing for a long time. And it's this idea that you have projects that you've accumulated throughout different companies, and you have to go through and explain what you did, how you did it, what your thinking was, how you influenced an outcome, how you shaped a decision, and things like that. And what I kind of realized is that in a world where anyone can basically fake evidence of something, a website, a deck, a paper, what they really want to know is like your true thinking. your true taste and how you leverage these tools tools to bring something to life, not just that you are capable of outsourcing your thinking.
Peter Shea: Yeah. They want to know how you got there. It's like Matthew just saying, I know I want to know how you solve the problem.
Kyle Bylin: Exactly right. I'd be curious, we're starting to see the rollout of these 24-7, 365 agents from Meta and OpenAI. And while they're currently on the professional tiers of these services for OpenAI, Meta's is freely available to the public. What role do you think agents are going to play in the classroom the near-term future? Do you foresee a world where students starting up as freshmen? might be assigned their own agent to help them manage their academic journey.
Peter Shea: Yeah, I I I think some universities have already begun exploring the idea of having an agent follow you from freshman year to graduation. I think that's certainly one potential use of agents in higher ed.
Kyle Bylin: Absolutely. I'd be curious, let's imagine a student arriving at Middlesex in twenty thirty. What would your hope be for what their education looks like? What is one future that you're excited to build?
Peter Shea: I am would be excited to build a educational system that does a better job of bridging the academic learning experience with the workforce preparation. I think there's always been a gap between those two spaces and we've expected students to jump over the gap, like a middle like the middle portion of a bridge that was missing. I think the most exciting thing is if we get this right, we will have finally provided the middle part of the bridge where we can transition students from the academic learning. into workforce learning and perhaps even back and forth, you know, traveling back and forth as needed.
Kyle Bylin: In the presentation that you shared with me, you recommended a book called Reshuffle. I'd be curious, what was the idea from that book that has been most useful for you in terms of how you think about education and work and how AI could change the labor market?
Peter Shea: I think the interesting thing about reshuffle is that it anticipates one of the ways in which this technology is likely to change the workforce, which is to say it's going if a job is a collection of tasks and a task half of them can be done by an AI, the rationale for that traditional arrangement of responsibilities in one job no goes away. You either make it a part-time job or you take those human-centric tasks and repackage them with some other role. So we be we need to be mindful that. Not only will new roles emerge, but repackaging of existing jobs will begin to emerge and students should begin to think about that in order to prepare themselves for the work groups.
Kyle Bylin: Absolutely. And it it's also this idea that like a job is really just a collection of projects and tasks that exist in an organization. And increasingly, more of those tasks are being automated. And as these agents gain more autonomy and more capability to work on longer time horizons, they're beginning to really work at a level of Doing complex projects in a short amount of time. One example I have is I was working with Chat GPT work, and I said, you know, I've got this interesting idea for an AI conference. Could you put together the website, the presentation, the programming, and any other assets that I might need, like a poster? And of course, I gave it lots of context for what I wanted. I talked about this model of like the coming wave and how each succession of the wave could talk about evolutions in AI, going from chatbots to agents to general intelligence. And it must have worked for 15, 20 minutes. And all of a sudden, I was looking at a fully fleshed-out poster, a fully designed website. A 20-slide deck that kind of walked through the idea and the budget. And all of a sudden, really made the starting point something totally different. It's almost like the orchestration of this agentic ability. It's not that you can outsource your thinking and all of a sudden get an AI slop conference. It's that if you really do have enough experience. and you have an idea that is big, you can now have AI help you bring those ideas to life on a time frame that would have blown our minds even four years ago, but you know, is is maybe seen as a small miracle today.
Peter Shea: Yeah, it is a miracle. It is a miracle. And one day it won't be seen as a miracle, which in itself is miraculous. it'll be as generation being born today will never know a world without AI. that's interesting to contemplate, you know.
Kyle Bylin: We've been talking a lot about kind of the current state of AI in higher education and the ways that agents and chatbots are a part of the classroom. I'd be curious when you kind of look at the horizon as a futurist and you try to communicate the change and the progress that is happening. One, what future do you envision? And two, Do you ever have challenges in communicating that vision in a way that gets buy-in from different stakeholders instead of pushback?
Peter Shea: Well, I fortunately I I I was intri interested in innovation and tech and education years ago. and I started out with as an unabashed enthusiast, just sharing it my enthusiasm and thinking that would be contagious. And then an older person, a manager, took me aside and in a very caring auntish sort of way, said to me, Peter, you and your friends scare people. You talk in a way that you amongst yourselves, like a bunch of D and D enthusiasts, you totally understand. People outside that sphere don't get it. And this things you talk about often makes them feel dumb without you realizing it. So you gotta find what if you're gonna really advance change, you have to think about your audience carefully, their concerns, their interests, their incentives, their fears, and then shape your your your message to meet that audience. And I think after thirty
Kyle Bylin: Absolutely.
Peter Shea: years of being in higher ed, I I've a I know I know my colleague's mindset. And I think I can speak to it.
Kyle Bylin: And how would you describe that mindset more broadly?
Peter Shea: I think for the mass majority of educators, they are inherently conservative in terms of their practices. I like to say that many college professors, while they're politically progressive, they're pedagogically conservative.
Kyle Bylin: Mm-hmm.
Peter Shea: They don't really want to change their own world. And you have to realize that because I think higher education is an industry that tends to attract people who are very change adverse. Regardless of their interest in students and regardless of their interest in scholarship, they like the fact this is a world that seems unchanging. least the outside and at least comparison to other industries. And so that makes it unusually difficult to do any kind of effective change management.
Kyle Bylin: You know, I think it's interesting that change management is the subject that often gets brought up, and it's one that I've started coming back to myself. Because this isn't like just a typical project where you can just go out and do it and it'll work, and everyone's just on your side. You're managing a complex situation, lots of different stakeholders, lots of different levels of AI literacy. You're working amongst an institution that Is silos among silos among silos. And what you're attempting to do as a practitioner is to essentially make change happen. And that's not an easy thing. What do you think have been some of your greatest lessons in change management in the past few years as you've attempted to bring AI into education?
Peter Shea: Know first what people care about, not what you think they should care about, but what they actually care about. Listen very carefully. Sometimes the things they care about are not the things they're talking about. And you have to glean from you have to understand you to keep be very alert. This is the counselor role. You have to be alert to the subtext of their remarks. I think that's very important in being being an effective communicator in this sphere. and and and you have to address that when you shape your message. So, yeah, you know, it's like anything else. What are the incentives? What are the disincentives? Be mindful of that. Be mindful of timing. Be mindful of the right measures at the wrong moment won't deliver. You've got to have a sense of when the moment opportunity comes, and then you know. you can't force change management in higher education. It's it's not a good idea anywhere, but you have to be the worst thing is to force it too early, and the other thing is to miss the moment because you're being too cautious. So you have to be alert to the opportunity. And then when it comes, be ready to move in. I was a fifty years ago, I was a Cub Scout. Be prepared.
Kyle Bylin: Absolutely. We've covered a lot on today's show. I'd be curious, is there any question that you wish I would have asked or any kind of final insights you want to share with the audience about what your experience has been like or potential advice you might offer to someone who's engaged in this kind of work?
Peter Shea: Don't do it alone. Find your friends. Don't be isolated. If this can be a change in management, it can be scary, frustrating work and you'll burn out if you if you don't have people to talk to and who can share your vision and work together. Treat it as a community project rather than a one person.
Kyle Bylin: Well said. And Peter, if people want to learn more about the community college or your work in AI, what are the best ways to find you online?
Peter Shea: I'm on LinkedIn. I have an AI and education group I started in twenty twenty. and they can simply reach out to me via my college email. I'll I'm always open to chat with people, particularly colleagues in higher ed who are trying to figure out how to navigate this at their own institutions.
Kyle Bylin: Well said. Well, thank you so much for coming on the show today.
Peter Shea: Thank you for having me, Kyle. I appreciate it.
Peter Shea: Thank you, Carl. It's a pleasure to be here.
Kyle Bylin: Peter, to start out, could you tell me about your role at Middlesex? What does the assistant dean for AI integration actually do? And what are you ultimately hoping to accomplish across the college?
Peter Shea: I think my role in has three different functions. One is to help lowering the anxiety around AI as an as a enormous technological and cultural change. Two help faculty, staff and students figure out what connection, if any, AI plays with their legacy practices and approaches. And three, help them envision How AI can be used to do entirely new things that we could never do before thanks to its unique affordances.
Kyle Bylin: And can you speak on the anxiety side of things? Obviously the pace of AI continues every day, and over the past four years it's certainly disrupted higher education a lot of different ways. What kind of anxiety are you seeing around campus?
Peter Shea: Well, you know, my campus, my college has always had a place for innovation. We were the first community college in Massachusetts to have an online program. And we were exploring things like immersive learning in the virtual worlds over a decade ago. So culturally, we you know, we weren't as shocked as other colleges were. We even played with a pre-generative AI tool around twenty eighteen or twenty nineteen. but what I am seeing in the larger Higher education space where I also operate as an advisor. I would say year one of AI's arrival was very anxiety-producing and disloc, you know, dislocating. Year two and three were still anxiety, but less so and a cautious engagement with AI. And this now, in what we are, I guess, year three or four. I'm sensing still engagement. People understand it's it's part of the world and we need to forward, but there's also I think more hostility towards AI than I I sensed in the past. because they're seeing the public narrative around AI has been associated with job loss, it's been associated with environmental impact of data centers, it's been associated with potential apocalyptic scenarios. AI is good at many many things, but it's not good at getting good press. and higher higher education is one of those fields where it has been harder to put to get a consistently positive relationship with AI. And despite a lot of effort, there is still considerable ambivalence, more ambivalence than I think in than in other fields.
Kyle Bylin: What do you think that the challenges are of overcoming this kind of bad PR problem that AI has? For people paying attention, as you're saying, the headlines are talking about job loss. The next week they're talking about, you know, existential risk to humanity. And meanwhile, we're all kind of still talking about a chat bot, we're talking about agents, but it seems like the potential of these capabilities can get lost in the conversation when we make it specifically about the environment, specifically about data centers, copyright infringement, and like all of the other kind of challenges that led to the AI that we have today.
Peter Shea: definitely. I you know, I I think one of the disadvantages for AI is that it's it's public public representatives, you know, in terms of the heads of the AI companies, are not the kind of people you would necessarily want. They're not a PR person's dream, to put it this way. They're not the best spokesman. There's no there's no Steve Jobs among the AI crowd. There's no one of of equal charisma. And a few of them have what you might call negative charisma. so
Kyle Bylin: Mm-hmm.
Peter Shea: also often the messenger is the part of the message. So that's been a bit of an issue. but there is a big you know, after all this negative feedback, there is a big there is some degree of growing awareness that AI has an enormous potential to to do more good than bad. In fact, the New York Times, I think it was On oct it was October second. they published a guest study called Six Charts That Show Just How Much We Need AI. And it addressed such things as we've gone through a productivity slump in the past few decades. We went through a lot of growth, economic growth, and then we just kind of like stalled. And we need some sort of mechanism to to to to reboot our both our economy and our you know our general culture. And AI is the kind of technology that has the possibility of doing that. So I I see it as as a as a great tool to to restart the the economic engine that you know sort of stole the nineteen seventies for the United States. and I think as a tool properly used can can increase people's sense of agency rather than diminish it. unfortunately, right now, we're we're still People are still using it like a sort of supercharged Google, which it's not, you know. I think it was is it David Sow of MIT who just published a book a few days ago where he talked about the excitement around the personal computer was because, as Steve Jobs said, it was a bicycle for the mind. So that was the bicycle for the mind. He said, AI is a motorcycle for the Which is to say it can get you a lot further than a bicycle, but there are dangers with a motorcycle that you don't have with a bicycle. so the I think it's a pretty good metaphor. I think we can get to places that we've never been able to get before. and I I'm very excited about the possibility. But unfortunately, particularly in higher ed, which is a inherently conservative industry, we are intent on looking back and simply preserving things, even things that have not been working out very well, because we're Attached to tradition rather than looking forward to what we can build that we can never build before. And I I think a lot of you know, my public communication around AI is trying to get people to look away from the preservation and more towards the creative construction. Or I like to say, stop, stop talking to the firemen and start talking to the contractors, you know.
Kyle Bylin: I think that's such a a rich metaphor, this idea that Steve Jobs said that the computer is like a bicycle for the mind. And I love your extension of that metaphor, that it that it's more like a motorcycle. What I would add on that is this idea that we all brought our creativity, our intelligence, our imagination to computers. And for For the vast majority of us, the keyboard was the main way that we could to talk to an intelligent machine, right? But these days, through voice prompting and obviously machine intelligence, now we can basically communicate in the with the computer through English, but then have that transformed into code and any other modality that we could potentially create. So in many ways, We are now the bottleneck in the sense that the limit of our imagination in terms of what we can envision is really the only limit as to what is possible right now. And it's interesting to kind of communicate that at a time when we're really focused on: well, is the model smart? Does it get everything right? As opposed to the human side and the empowerment that these new technologies unlock.
Peter Shea: Yeah, we are we are new, you know, whenever a new major technology emerges, we spend at least a decade negotiating our relationship with the technology. it's a it's an awkward dance. but we've been here before. People forget that. If you're familiar with the history of technological integration, the you know, things which seem so seamlessly a part of modern life were often resisted or misunderstood in the first few decades. And A good example would be the automobile. we tend to imagine that the minute Henry Ford opened his factory, everyone was buying cars, and that's not the case at all. It took a long time to shift from a horse based travel methodology method to an automobile centric one. And in many in many key industries, resistance to changing from horses to automobiles or mechanized vehicles was was fairly strong, particularly in the military, which had a lot of people, senior officers who were invested in the cavalry and the horses, and who did not want to see their expertise in that area become obsolete. So and I I think you see that a lot with college educated people and public intellectuals arguing against AI. Part of the thing AI threatens is there is that The value they often provide is their expertise, but also the scarcity of their expertise. And when their expertise becomes less scarce, it becomes less valuable. And so there's a threat to their own economic well-being, which I think manifests itself often in s and maybe not consciously, but it manifests itself a lot in the public criticism of AI. but I again I think you know. You can rely upon AI overmuch and allow it to weaken your natural capabilities. But you don't have to. You have a chance to define your relationship with the technology. You're not passive, you have agency. That's what people forget. You know, some people are gonna choose to use it lazily, but the night the same people use all technologies lazily. You can say no to certain practices and say yes to others. And for me as an educator, I'm more interested in AI as a tool that can augment human ability and develop human and develop the ability that then lets later gets augmented. So I see it not just as an extension of human ability, but as a tool properly used, which can give us the kind of immersive learning experiences that enable us to get really, really good at things and which are full potential. And an example of that I would be is that. Is we simulation-based learning. It used to be very difficult to build a meaningful simulation, which was unfortunate because they were uniquely powerful in in producing really high quality learning for people, particularly in demanding fields. I mean, that's what got us to the moon. That's three astronauts in a in a space simulator. But now with with AI, we can we can build and scale up simulation in a way we never could before. I think that's right. I think there are two things about AI. in education that have been overlooked in all the fury. And one is AI has effectively closed the door on a lot of practices which were flawed, but we continue to do so because people were attached to them. It's like Balboa burning his boats behind him. It's going anywhere, like there's no there's there's no way back. You can't we we can't claw claw our way back to that world no matter how many blue books you take out and have students write in them. The second thing is AI has potential to build learning experiences and scale them in a way it's been utterly impossible throughout history. And that I think is revolutionary. And that I think is getting overlooked in all the public furor about AI.
Kyle Bylin: I'd love to hear more about that last point that you made, which is that these educational experiences at scale are now possible through the different tools and services that these AI companies have made available to us. Can you talk about what your thoughts on that are, or maybe some examples from Middlesex that you guys have done in trying to roll out educational resources?
Peter Shea: Yeah, sure. One of the things I like about AI is that you don't you can build things in-house. I mean, I've been involved with educational technology for thirty years. And that, you know, apart from basic web pages, that almost inevitably required c contracting with a vendor. At Middlesex, we've built a lot of things in-house using an AI API key to to power it. So if you view AI as a form of energy, then you build the things you wanted to energize and you then you can experiment with them. And we built a number of AI powered tools in-house at very little expense that we can implement. And I think that it creates a new it creates a new dynamic for educational technology because before you were you were entirely dependent on vendors. You know, w one of the most exciting parts about the early stage of the internet was back when. Ordinary people were creating their own web pages and deploying them. They weren't having to pay someone to do it. That was a really liberating experience. And that's what unleashed a lot of the excitement about it. You didn't have to pay someone. I think with AI, now we have the same experience all over again. We we we don't have to we don't have to go to a vendor, we don't have to go to a company organization. We can build on our own. And I think that's enormously liberating. and I it's not only that, it's also very cost effective, which I think is another factor which is going to really help you know increase the demand for a AI-based tools. the problem I think is in higher education, it's we're trying to attach a an engine to a horse. and then when it doesn't work, we blame the engine. It's not the engine, it's the we have to have conversation about the horse. In fact, if I ever wrote an article about it, I probably call it we need We need to talk about the horse. Because a lot of people in education see themselves they're not they don't see themselves in being tra in as a as a comparison. They don't see themselves as being in the transportation business. They see themselves in the horse business. And I say, no, we're in the transportation business. We're trying to get people from here to there. And the horse has been traditionally the way we get there. And I know we've been doing it for a long time, but now there's a better method. And that's really it requires a real paradigm shift for people, which is obviously always difficult for anyone.
Kyle Bylin: In a presentation that you shared with me, you described three parts of your current job, which is counselor, facilitator, and futurist. I'd be curious which of those takes up the most time right now in terms of the capacities you kind of have to fulfill on campus in those different roles.
Peter Shea: I think in the beginning, counselor was a big part of it. And I knew it would be. But I am spending more and more time in the futurist role, which is what I really want wanted to be to begin with. but you have to lay the ground. I often say in in a job like mine, the absolute you don't you don't need to have a technology degree or background, but you need to have like an anthropology or psychology background because it's not the challenges aren't technological. Someone's already figured that part out. It's the human change management element that is essential. And you can know a lot about AI, but if you don't know much about change management or the culture of your organization, you're going to fail. So
Kyle Bylin: I I can absolutely see that. to pick up where you kind of left off there, I definitely agree that that some aspect of this requires us to be a bit of a futurist. And in fact, futures thinking is what I have integrated into my own AI lectures to students, which is that I I paint this picture that the frontier AI labs since their beginning. Have been working on this idea of AGI or artificial general intelligence. And some of them are even saying that either they've achieved it now or they might be close within a few years. You combine that with different predictions from the World Economic Forum's future jobs report and how the market could shake out by 2030. And you have a world that In the four years that a student spends with us, could be very different by the time that they graduate. And so I try to empower them with prompts and tools and concepts to where they can talk to AI about their specific field and get personalized recommendations about what opportunities are emerging, what careers are potentially going away. and how to essentially use AI to keep abreast of all the different news and events in their field. That way, if they pay attention to this, if they track the changes that are happening, AI can be something that potentially empowers them rather than something that they feel is happening to them.
Peter Shea: yeah, I I definitely think one of the things that colleges have to do is figure out how to talk about AI with their students. They have to have the talk and they had they they can be candid about their concerns, particularly in regards to the way that the misuse of AI can inter intervene with productive this productive struggle for l needed for learning. That's perfectly legitimate and absolutely required. But the same time they have to acknowledge that their students are going out into a world where it's going to be shaped by this technology. And they have to give them some sense of how to begin developing their own relationship with it. And this is a great opportunity for them to help shape that relationship positively rather than leaving students up on their own. And when they don't do that or they're just simply throwing up their hands, then they're I think they're failing in their obligation to students.
Kyle Bylin: Absolutely. And like building on this discussion, we've talked about this idea of potentially becoming over-reliant on AI, potentially outsourcing both your doing and your thinking. You started in this space exploring how AI could be applied to student writing and how it could give them feedback. I'd be curious, like looking back, what did you see then? And now how do you think about AI's involvement in the writing process today?
Peter Shea: I think AI properly used by both the faculty member and the student can can enormously enhance a person's ability to develop their writing skills. But again, it's it's how you it's not whether you use the tool or you don't use the tool, it's how you use the tool. Like a scalpel. I mean a scalpel can can be very dangerous in the wrong hands. you can hurt yourself. Same with a power drill. If you use them properly, you can do remarkable things. Students would have to understand what part of the writing process, particularly when they're learning, should not be outsourced to the AI and which which the student themselves have to own. On the other hand, an AI's ability to analyze a particular student's text and give them targeted feedback. Is uniquely valuable. I've never seen anything like that before in my life. And I've seen AI give better feedback to an example of student running than than I've given because the AI has such unique analytical powers. And feedback is crucial to developing people as writers and communicators. So I I I see it as a largely positive relationship.
Kyle Bylin: It's been really interesting to watch the idea of writing evolve as the models have gained in their capacity. I remember sitting at the research desk next to another librarian about two years ago, and they talked about this idea that AI could could fabricate a paper essentially, and it could also come up with fabricated citations. And at the time, You know, this was seen as like a limit on the system, as if it would always be there. And as you could guess where this story is going, a few months later, deep research came out. The AIs got better at doing research, synthesizing information together, and producing citations. And since then, I believe it has got gotten even better and potentially even more accurate. To where, if you look at that study done by MIT, they suggest that most undergraduate homework at this point could potentially be automated. And I think that like raises a question in terms of what exactly is the homework that's being assigned to students? What are the outcomes that we're trying to achieve? And are these the same outcomes that they're ultimately going to be asked to create in the workforce, essentially?
Peter Shea: yeah, you know, again, the the idea of yeah, the idea of homework itself is is gonna be challenged, particularly for K through twelve. I mean if I had a K through twelve student, I I wouldn't I would not I wanna I would want some way to track my child's progress, but I wouldn't want I ho in my experience homework was often assigned as a way of reassuring parents, as a way of so they can see what their stu their kid is doing. rather than just relying upon you know report card grades. my view is that if you know if you're spending six hours a day, five days a week in school, there's really no reason to be taking work home with you. You I mean you should be able to you should be able to do all your productive learning work at school. And when you're not in school, you should that should be your time to spend. So I'm not a big I'm not really mourning the death of homework. and obviously it's different for college students. I just think we have to The other thing that we have to deal with is is is figuring out new ways to assess students' thinking and skill development that rather than our our legacy because the attachment to essays and written exams was it was perceived as the most practical way of of assessing students' knowledge growth. that's no longer the case. We need new assessment tools. And so that's going to be a big, very big conversation in higher education because you know, we like to say that we're in the education business, but we're really in the credentialing business. We we sell credentials. People can people can learn things without us, but they can't get a credential without us, at least not in most fields. But if our credentialing system is broken and everybody knows that our business model is is you know under threat. So I think you know, the question is really: how do we are there better ways to assess? what students are learning and and how they can apply it.
Kyle Bylin: I'd be curious what have been your takeaway so far. I, like you, went to the teaching and learning conference with AI in Florida. And it was interesting to see all of the different ways that people are incorporating chatbots and maybe early stages of agents into the classroom. But ultimately, the assessment of the work was still kind of a challenge. Because you were trying to figure out at what point do you assess the students' work and what kind of like work are you attempting to s to assess? So I'd be curious like if you learned anything from that conference in terms of innovative assessment and to what extent you guys have been able to apply it at middle sex in any of your different departments.
Peter Shea: Well I think you know In my presentation at the conference in Orlando, I did an analysis of all the proposals and I noticed that a lot of them were around building frameworks, which I thought was interesting. You know, frameworks is our version of of a of a of a dike. You know, we're trying to keep the water from coming in. But you you can only get the dike so high up. At some point, you have to realize that you gotta retreat and and and stand your ground somewhere else. at Middlesex, for example, we explore using simulation tools to capture assessment data so we can track people's decisions over time. It's much harder to cheat in in a simulation-based assessment than than in conventional ones, and it can be it it can render much more interesting data, I think. I you know, I've always been struck by the potential of learning analytics. but I don't think learning analytics ever achieved its full potential because there were no ways to really extract interesting data using our conventional methods. I mean I used to go to presentations where people would get excited to say, you know, we're tracking when students log into their LMS and we're tracking when they log off. And it looks like many of them are only looking at the assignment three hours before it's due. And I remember one group of presenters that were so pleased by that. And I'm thinking this is the best you can do in terms of the data you can sift. But if we build tools that track decisions and actions, we can get much more interesting data to to assess student learning than we ever did before. And that's the kind thing that excites me.
Kyle Bylin: Absolutely. Zooming out a little bit, I'd be curious to know how would you describe the way that middle sex has kind of responded to this AI moment? Where do you believe you guys are making progress versus what are the areas where you feel like you're still trying to play catch up?
Peter Shea: Well, I don't think there are too many places we have to worry ketchup because I don't think there are many peop people in the lead to begin with. we're still kinda all absolute beginners. I think that the fact that there's a full that I have a full time role doing this says a lot about where the college is at and the fact that we have a three year strategic plan. And we just recently acquired a a Title Three grant around AI integration to support that. So I think that helps establish, you know, where we are on the field.
Kyle Bylin: And when you kind of zoomed out to the Teaching with AI conference and you saw what everyone else is doing, what was your kind of analysis of the overall field? I certainly felt like there's a lot of pilots, there's a lot of interesting spaghetti being thrown against the wall, but no one really knows what is sticking. And because we have to deal with, you know, faculty and students in real time, it doesn't necessarily always guarantee buy-in. And one year you could be all in everyone's favor and have people attending your various events and stuff like that, but the next semester it could be a total ghost town and you can run into, you know, trouble sustaining that same momentum.
Peter Shea: That's why I think those of us who are at the frontier of trying to advance this work need to speak with one another on a regular basis and support one another as a community of practice because we all have the same issues. And the buy-in is going to fluctuate and it's gonna be different from institution to institution. I think a a small I think a thousand campfires is hard to sustain. I think there needs to be more coordination, more sharing of innovative practices that are that are replicated and can be scaled up. I think we need to have some sort of association where we're we're we're trying new things and a having a shared experimental space and when we come up with something that seems to work, share it out with the others so that we you we can all invest in one another's success.
Kyle Bylin: I'd be curious as you look across the landscape of higher education and the different AI initiatives that you've seen roll out, are there any that you felt were especially on point and that you wished you could, you know, bring to your own campus?
Peter Shea: I'm I'm very impressed by what they did down at Miami Dade College in terms of the creation of NIAC, the National AI Consortium, where they immediately recognized the need to help design courses around the AI fluencies that certain employers would want. I think they really took the bull by the horns and and you know, had a vision and an operational plan. And I'd like to see some of my initiatives flourish the way theirs have. I think they've been a really they've been a really great leader in this field.
Kyle Bylin: love to hear more about the initiatives that you're trying to do at Middlesex. What are the bright spots in terms of successes that you have or things underway this semester that you feel like are really going to help get things off the ground?
Peter Shea: I like using AI to tackle bottleneck problems. you know, one of the ways in which we're looking at it, having AI review writing samples and then giving the students advice, not not telling them what to do, but giving them advice as to what section of English they should probably enroll in when they come to the college. the volume of of of Writing samples that we see every semester is is is a lot and and you know, the process of evaluating them can be slowed down and AI can really help address that in a speedy way. That's one example. Another example would be a tool we had that takes a job description and then identifies potential questions the applicant might get during an interview. I think for our students that's that kind of is very valuable because many of them are not at the point where they can envision what they're gonna be asked in a job interview. So I think using AI to solve problems that have seemed unsolvable before, I think is really one of the best applications of it.
Kyle Bylin: I I definitely agree. I think this concept of the wicked problem is something that aligns itself well with AI, just in the sense that if we look out to our communities and the broader world, there's all types of wicked problems, whether it's like human poverty and food access to even the higher levels of climate change and management. And What we need more than anything is people who can use their domain knowledge, their unique taste and agency to leverage these tools to potentially find answers or solutions to some of these wicked problems that perhaps we may not be able to do on our own, or now suddenly small teams of students can make tremendous effort towards even in a semester, let alone during their entire undergraduate curriculum.
Peter Shea: you know, I I think I don't think most colleges are going to I don't think most colleges are going to do radical changes to the curriculum. I don't think they're equipped to do that. I think they have to be adaptive of it. And I think some colleges will become very fluid in adaptation and curricular change. One thing that's gotta change is curricular view and adoption has to move quicker than it has in the past because I I recall I recall talking with someone who once had the job titled Provost for Innovation and then he left that job and went back to being a professor. And I said, Well, why is a great title? He said, No, it's not. He said, We we we we would propose a course that we really saw a need for, and it would take three years to get it through curriculum review. So I think things like that. I the the the structure of of programs and disciplines may not change, but I think the curricular review process will change. So that's the one at least my my prediction.
Kyle Bylin: I could definitely see that as being a challenge. It's ones that I've ultimately seen in my own work too. I'd be curious, something that I've wondered about for the past couple of years is this idea that we had the divid digital divide, which talked about access to computers and the internet and all of the opportunities that the digital age could unlock. But now increasingly I think there is an AI divide emerging. Not just in colleges, but also in curriculums, in the sense that some institutions are providing full access to paid tiers, whereas others have students that are ultimately footing their own bill. But beyond that, you know, access to the intelligence and the paid tiers, I also think is like the cultivation of the mindset of what is possible. And what I mean by that is that there's just Something distinctly different from like learning what AI is, how a large language model functions, and basic prompting capabilities when compared with some of the education that is emerging on campuses like Stanford, for example, where I can imagine that vibe coding was established very early on. Students are prototyping with these tools, they're designing products and services. Rather than kind of being worried about this generation of information and how it can be used for plagiarism, they're talking about the use of code and how it can be used to build businesses that potentially further the lives of students and maybe create opportunities for them that that weren't even there before until AI democratized coding itself to everyone.
Peter Shea: Yeah. And what's more, in certain professions, the ability to work with AI is going to become essential. Now it's not going to be equal across the board, but I've heard from like the CEOs of IT companies that have said one of it bluntly to to me and a group of others, if your computer science students don't use AI, don't bother to send them to our job interview.
Kyle Bylin: Absolutely. And like the reality is, like, I feel like that is becoming more of the norm in a previous conversation that I had with Jessica Stansbury. she talked about this idea that the AI portfolio is increasingly becoming a thing that employers are looking for. And I thought that was really interesting because in the user experience design and research profession, portfolios have been a thing for a long time. And it's this idea that you have projects that you've accumulated throughout different companies, and you have to go through and explain what you did, how you did it, what your thinking was, how you influenced an outcome, how you shaped a decision, and things like that. And what I kind of realized is that in a world where anyone can basically fake evidence of something, a website, a deck, a paper, what they really want to know is like your true thinking. your true taste and how you leverage these tools tools to bring something to life, not just that you are capable of outsourcing your thinking.
Peter Shea: Yeah. They want to know how you got there. It's like Matthew just saying, I know I want to know how you solve the problem.
Kyle Bylin: Exactly right. I'd be curious, we're starting to see the rollout of these 24-7, 365 agents from Meta and OpenAI. And while they're currently on the professional tiers of these services for OpenAI, Meta's is freely available to the public. What role do you think agents are going to play in the classroom the near-term future? Do you foresee a world where students starting up as freshmen? might be assigned their own agent to help them manage their academic journey.
Peter Shea: Yeah, I I I think some universities have already begun exploring the idea of having an agent follow you from freshman year to graduation. I think that's certainly one potential use of agents in higher ed.
Kyle Bylin: Absolutely. I'd be curious, let's imagine a student arriving at Middlesex in twenty thirty. What would your hope be for what their education looks like? What is one future that you're excited to build?
Peter Shea: I am would be excited to build a educational system that does a better job of bridging the academic learning experience with the workforce preparation. I think there's always been a gap between those two spaces and we've expected students to jump over the gap, like a middle like the middle portion of a bridge that was missing. I think the most exciting thing is if we get this right, we will have finally provided the middle part of the bridge where we can transition students from the academic learning. into workforce learning and perhaps even back and forth, you know, traveling back and forth as needed.
Kyle Bylin: In the presentation that you shared with me, you recommended a book called Reshuffle. I'd be curious, what was the idea from that book that has been most useful for you in terms of how you think about education and work and how AI could change the labor market?
Peter Shea: I think the interesting thing about reshuffle is that it anticipates one of the ways in which this technology is likely to change the workforce, which is to say it's going if a job is a collection of tasks and a task half of them can be done by an AI, the rationale for that traditional arrangement of responsibilities in one job no goes away. You either make it a part-time job or you take those human-centric tasks and repackage them with some other role. So we be we need to be mindful that. Not only will new roles emerge, but repackaging of existing jobs will begin to emerge and students should begin to think about that in order to prepare themselves for the work groups.
Kyle Bylin: Absolutely. And it it's also this idea that like a job is really just a collection of projects and tasks that exist in an organization. And increasingly, more of those tasks are being automated. And as these agents gain more autonomy and more capability to work on longer time horizons, they're beginning to really work at a level of Doing complex projects in a short amount of time. One example I have is I was working with Chat GPT work, and I said, you know, I've got this interesting idea for an AI conference. Could you put together the website, the presentation, the programming, and any other assets that I might need, like a poster? And of course, I gave it lots of context for what I wanted. I talked about this model of like the coming wave and how each succession of the wave could talk about evolutions in AI, going from chatbots to agents to general intelligence. And it must have worked for 15, 20 minutes. And all of a sudden, I was looking at a fully fleshed-out poster, a fully designed website. A 20-slide deck that kind of walked through the idea and the budget. And all of a sudden, really made the starting point something totally different. It's almost like the orchestration of this agentic ability. It's not that you can outsource your thinking and all of a sudden get an AI slop conference. It's that if you really do have enough experience. and you have an idea that is big, you can now have AI help you bring those ideas to life on a time frame that would have blown our minds even four years ago, but you know, is is maybe seen as a small miracle today.
Peter Shea: Yeah, it is a miracle. It is a miracle. And one day it won't be seen as a miracle, which in itself is miraculous. it'll be as generation being born today will never know a world without AI. that's interesting to contemplate, you know.
Kyle Bylin: We've been talking a lot about kind of the current state of AI in higher education and the ways that agents and chatbots are a part of the classroom. I'd be curious when you kind of look at the horizon as a futurist and you try to communicate the change and the progress that is happening. One, what future do you envision? And two, Do you ever have challenges in communicating that vision in a way that gets buy-in from different stakeholders instead of pushback?
Peter Shea: Well, I fortunately I I I was intri interested in innovation and tech and education years ago. and I started out with as an unabashed enthusiast, just sharing it my enthusiasm and thinking that would be contagious. And then an older person, a manager, took me aside and in a very caring auntish sort of way, said to me, Peter, you and your friends scare people. You talk in a way that you amongst yourselves, like a bunch of D and D enthusiasts, you totally understand. People outside that sphere don't get it. And this things you talk about often makes them feel dumb without you realizing it. So you gotta find what if you're gonna really advance change, you have to think about your audience carefully, their concerns, their interests, their incentives, their fears, and then shape your your your message to meet that audience. And I think after thirty
Kyle Bylin: Absolutely.
Peter Shea: years of being in higher ed, I I've a I know I know my colleague's mindset. And I think I can speak to it.
Kyle Bylin: And how would you describe that mindset more broadly?
Peter Shea: I think for the mass majority of educators, they are inherently conservative in terms of their practices. I like to say that many college professors, while they're politically progressive, they're pedagogically conservative.
Kyle Bylin: Mm-hmm.
Peter Shea: They don't really want to change their own world. And you have to realize that because I think higher education is an industry that tends to attract people who are very change adverse. Regardless of their interest in students and regardless of their interest in scholarship, they like the fact this is a world that seems unchanging. least the outside and at least comparison to other industries. And so that makes it unusually difficult to do any kind of effective change management.
Kyle Bylin: You know, I think it's interesting that change management is the subject that often gets brought up, and it's one that I've started coming back to myself. Because this isn't like just a typical project where you can just go out and do it and it'll work, and everyone's just on your side. You're managing a complex situation, lots of different stakeholders, lots of different levels of AI literacy. You're working amongst an institution that Is silos among silos among silos. And what you're attempting to do as a practitioner is to essentially make change happen. And that's not an easy thing. What do you think have been some of your greatest lessons in change management in the past few years as you've attempted to bring AI into education?
Peter Shea: Know first what people care about, not what you think they should care about, but what they actually care about. Listen very carefully. Sometimes the things they care about are not the things they're talking about. And you have to glean from you have to understand you to keep be very alert. This is the counselor role. You have to be alert to the subtext of their remarks. I think that's very important in being being an effective communicator in this sphere. and and and you have to address that when you shape your message. So, yeah, you know, it's like anything else. What are the incentives? What are the disincentives? Be mindful of that. Be mindful of timing. Be mindful of the right measures at the wrong moment won't deliver. You've got to have a sense of when the moment opportunity comes, and then you know. you can't force change management in higher education. It's it's not a good idea anywhere, but you have to be the worst thing is to force it too early, and the other thing is to miss the moment because you're being too cautious. So you have to be alert to the opportunity. And then when it comes, be ready to move in. I was a fifty years ago, I was a Cub Scout. Be prepared.
Kyle Bylin: Absolutely. We've covered a lot on today's show. I'd be curious, is there any question that you wish I would have asked or any kind of final insights you want to share with the audience about what your experience has been like or potential advice you might offer to someone who's engaged in this kind of work?
Peter Shea: Don't do it alone. Find your friends. Don't be isolated. If this can be a change in management, it can be scary, frustrating work and you'll burn out if you if you don't have people to talk to and who can share your vision and work together. Treat it as a community project rather than a one person.
Kyle Bylin: Well said. And Peter, if people want to learn more about the community college or your work in AI, what are the best ways to find you online?
Peter Shea: I'm on LinkedIn. I have an AI and education group I started in twenty twenty. and they can simply reach out to me via my college email. I'll I'm always open to chat with people, particularly colleagues in higher ed who are trying to figure out how to navigate this at their own institutions.
Kyle Bylin: Well said. Well, thank you so much for coming on the show today.
Peter Shea: Thank you for having me, Kyle. I appreciate it.
