A Discussion on AI in College Admissions

How should AI fit into the college admissions process? What guidance are institutions offering prospective students? And, how can high school counselors advise their students? In this episode, we speak with Emily Pacheco, Advisor for AI in College Admissions. We talk about ethical use of AI for college admissions, the opportunities ahead and the need to establish “rules of the road” for AI. To learn more about AI in College Admissions, visit edhub.ai for more resources/events.

Transcript

Matthew Sterenberg (00:02.147)

All right, Emily, welcome to the podcast.

Emily Pacheco (00:06.69)

Thank you, it’s great to be here.

Matthew Sterenberg (00:08.773)

So I wanted to bring you on because I think the topic of ethical AI use in college admissions is fascinating and you’re part of an organization that is dedicated to this. Tell us a little bit more, like what do we even mean by ethical? What are we talking about with AI in college admissions? You’re a college admissions professional. Give us the lay of the land.

Emily Pacheco (00:33.198)

Well, you know, those questions you just asked are the questions this group is exploring. And my interest in this topic, you know, came out of finding AI extremely fascinating, finding interesting uses for it in my daily work. So I dug in and looked for the people that were doing things with AI in college admissions, because that’s my professional area. And I quickly found that there was a pretty big void of information. just… weren’t a lot of experts out there. They’re just, in general, this is a pretty new topic, so there aren’t a lot of people talking about it yet, is what I found about a year ago. So that was the birth of this group. I created this community to start looking at the effects of artificial intelligence on college admissions. And of course, in that exploration, I think… you know, the most important thing is that we’re looking at the ethical implications of this technology. So, you know, I think there’s amazing potential in what it can do, and that’s what I’ve seen, but I’ve also know it’s really important that the educators are in the room talking about this and looking at what it means for this kind of technology to start kind of infiltrating college admissions and… What we find is that students are the first to use this, you know, and students have been using this since the first, you know, day it was released. But educators are a little bit behind. And so I created this community where we’re just over 1500 people that are looking at kind of what we as educators can do to try to guide this conversation to be more a part of the conversation.

Matthew Sterenberg (02:19.109)

So let’s make this kind of real for people. And this always helps me as I understand a topic, like what are the major uses of AI or the potential uses of AI in college admissions? Where my mind goes immediately is the essay, you know, but that’s just one piece. Like, what are we really talking about when we say AI in college admissions?

Emily Pacheco (02:41.762)

Yeah, I think there are definitely kind of two sides to usage and it depends who you are in terms of how you’re looking at how it’s being used. There is the university institutional side. So that’s where admission counselors are using it. They’re using it to craft personalized emails, marketing campaigns to figure out which students they should reach out to. So there’s a lot that admission offices are doing with artificial intelligence to, change how they’re marketing, to change how they’re interacting with prospective students. that is also in the form of chatbots. So a lot of students have started interacting with pretty high functioning chatbots. These aren’t the old school chatbots that weren’t able to answer questions. These are actually pretty cool features that some universities like NYU has a really cool one that are answering pretty complex questions and helping students. I think the response there is a lot of times people are like, why would a student want to talk to a robot? They want to talk to an admission counselor. And it’s like, that’s great. We’re there from 9 a.m. to 4 p.m. And it can be hard to even during those hours to get somebody on the phone. You’ve got to make an appointment. You make appointment for next week. You get a 30 minute slot. And that is a big ask, I think, for a lot of 17 year olds. And the nice thing about a chat bot is it’s 24/7. Kids at 10 p.m. can actually interact, get questions answered, explore a university. there’s that side, the way that universities are using this technology to interact with prospective applicants. The other side of it, of course, is the side that you mentioned. And I think that’s the side that people think of usually first when they’re thinking about how AI is being used in college admissions. And I do think the first thing that people think of is the college essay. And actually that’s where I started in this. I started presenting last year on how college counselors could use artificial intelligence in an ethical way to help students in the college essay writing process. Because that has historically been something that takes a lot of time when a high school counselor wants to work with a student and help them really in a systematic way. And there’s a lot of different steps in that process that I’ve seen where artificial intelligence can help. A human counselor work with a high school student. We also know that students are using it in what most of us would consider unethical ways. And that would be, you know, putting in a college essay prompt and saying, create my essay. So asking chat GPT, you know, to fully form a college essay that they are then going to use. But I think anybody that has done that, if you’ve put a generic prompt into chat GPT, what you get back is something pretty generic as well. And so a lot of educators, you know, had that experience where they’re like, I’ve tried chat GPT, I’ve put in, you know, a prompt and it’s terrible. It’s absolutely awful. What came back was completely useless. And what, you know, what we find is, you know, if you put in crap, you get out crap. so, excuse my language, but yeah, when you put in something that’s not very good, what you’re going to get back is not very good. So, so we’ve done a lot, you know, helping educators kind of understand that students are using this in pretty advanced ways that we are not detecting. And they are using it in the essay writing process, both unethically, but also in really creative, cool ways to help them in that writing process. And I think, you know, that’s what we’re hoping to focus on with our community, because our community being the educators is to teach them to view this not as just a mechanism for cheating, but look at this as something that could help you in the way that you’re working with students. you know, whether that be either side of the desk, if you’re a university looking at how can you do this better? Like, you know, how can you market to students that are really interested in your institution? You know, not just throwing something out there for every student, but looking for the students that are really the right kind of student for your institution. And then the other side, using it to how do we help students in that college admissions process? It’s an extremely confusing process that has just gotten more less transparent and more confusing over the years. And I do feel that there is a glimmer of hope that these tools could be, you know, could help streamline that process and help us in the way that we’re working with students to help them get to university.

Matthew Sterenberg (07:18.627)

Yeah, there’s, there’s a lot of tension related to this. Like I, I just sit back and I go, okay, we don’t want it to be, we want the students own words to some degree in the essay, let’s say. And, but at the same point, like you said, an admissions office is going to use AI tools to like write emails, to market to students. You might even be using AI to analyze the essay that you get back. So it’s kind of this funny thing where it’s like, we don’t want you to use AI as students. We’re going to be using it in admissions to do all these different things, to write the email to you, to market to you, to even analyze. But it makes sense in the sense that we want to get to know the student. Are they a good fit for our institution? And we don’t want just something that you were able to get out of AI. We want it to be your own words. We want to get to know you. So I think you’re right. We have to have some rules of the game, but it’s so new. It is scary. there’s a few ways that this can go, in my opinion. One is people are like, well, we’re going to do away with the essay. There’s got to be some other way that we can get to know the student. And AI might even be part of that. How do you upload all of your achievements and we’re able to get a better understanding of who you are as a student. So I think there’s some interesting use cases that actually allow us to get to know the students better. But there are definitely things where I’m like, this is, you know, it’s too easy. If you know how to prompt engineer this college essay, you can come up with something actually pretty, pretty much it’s going to sound like your own words. So it is a very difficult thing to weed through because you have one group that really wants to use AI for efficiency. And they’re telling another group, well, don’t really want you to use it some, not all. It’s an interesting place to be.

Emily Pacheco (09:23.118)

Well, I think we have to be careful saying that universities don’t want students to use it. We don’t yet, we’re not making it clear how we want them to use it. And that’s a big problem right now. And I think I’ve heard a lot of people say that, that line of like, universities are using it, but they don’t want students to use it. And when you talk to admission people, the ones that know about how artificial intelligence works and all of the amazing things it can do, I don’t hear any of them saying we don’t want students to use it.

Matthew Sterenberg (09:31.973)

Mmm.

Emily Pacheco (09:52.662)

Actually, they say, we know students need to know how to use this and we know there’s really good ways that this can be used. So it would be stupid for students not to use this in the college admissions process, but we need to make it more clear what those expectations are. And right now we’re not doing that. Universities are kind of letting students decide that on their own because there aren’t policies, there really aren’t many policies that are specific to college admissions. In terms of providing guidelines and saying, this is how you should use it, this is how we expect you not to use it. And exactly that, like just creating the essay out of just entering the prompt and getting that. But there are a lot of really good ways that it can be used. And I do think universities aren’t against those ethical ways, but we’re behind in creating those guidelines. And that’s one of the things that the community that I’m a part of is working on is helping universities do that better. I have a document that’s collecting all of the policies specific to college admissions. Right now I only have eight universities on that that have put out guidelines like that. And I expect to see that number double, triple, and then just a floodwater of universities coming out with clearer policies. And I think that’s gonna happen in the next few years, but we’re definitely at this point that’s we’re leaving students really not knowing how to, you know, what to do in that process in terms of using artificial intelligence. And that’s really unfair. It’s unfair to them. It’s really, you know, makes it really hard for students at this moment.

Matthew Sterenberg (11:31.543)

Yeah, we need to give them the rules of the road, right? What’s acceptable, what’s not. You mentioned you have a few universities that do have clear expectations. What, what are they saying to students?

Emily Pacheco (11:43.767)

Yeah, there’s a number of schools that have come out. The best guidelines definitely give specific examples. They say, know, do this and don’t do that. You know, a lot of them are just a little bit more general. know, one of the guidelines that I absolutely love at the end, I believe it’s Georgia Tech has this at the end of their guidelines is, you know, at the end of all of, you know, reading these guidelines, if you’re still unsure if the task at hand, could be, should be accomplished by artificial intelligence, ask yourself if this is something that you would ask a trusted adult like a college counselor or a parent to help you with. And I liked that kind of, it’s impossible to lay out every guideline of do this, don’t do that. But it’s nice to say, think about this. Is this something that your college counselor would do for you? If you went to them and said, help me with this task, would they say, no way, you need to do that, that obviously is something you need to take, you need to take authority in that, or is this something that they would step in and say, yeah, we can help.

Matthew Sterenberg (12:51.747)

Right, you wouldn’t ask your college counselor to rewrite your paper or something or rewrite your essay, but I remember doing this, I still do it, where you go to thesaurus.com, you’re like, what are some other words I can use here? What’s a different way to phrase this, right? That to me, of course that’s acceptable, right? It’s just another resource of how do I translate my own thoughts and words into something that’s more compelling or tells a better story or something like that.

Emily Pacheco (13:18.156)

Yeah, yeah. And unfortunately, that’s not black and white, we find. mean, you to me, you’d say, yes, that makes complete sense. But recently I was talking with a bunch of counselors and I used the example of helping a student brainstorm a title. You know, and to me, that was something like when I was, I still remember going to my mom with a paper and saying, I need a good title, you know, what would be a good title? And she’d help me brainstorm that. But in this room, there were several college counselors that said, no, using AI to brainstorm a title. You know, is overstepping that line. So it’s very clear to me as I navigate in this, you know, and talk to counselors, this is not black and white. There’s not, there’s very little within this that you can say like everybody would agree on this or that. It’s actually something that’s, that’s pretty nuanced. And I think what we’re going to see is it’s something that is just going to be evolving over time. So the policies that we need to make, we actually need to make them flexible. We need to be able to kind of

Matthew Sterenberg (13:50.351)

Mmm.

Emily Pacheco (14:15.598)

come back to them and say, okay, this made sense two years ago, but, you know, things are changing and that policy is going to change as well.

Matthew Sterenberg (14:21.933)

Yeah. And AI, mean, how good is it going to be in two years? Like we, it’s good now. Like the idea that we are not going to have to revisit this constantly is it’s evolving so quickly that I think you’re going to have to consistently revisit it because it’s, we have no idea how good it can be, which is the scary element of all this, right? Like you could put your, you could be

Emily Pacheco (14:48.322)

Yeah.

Matthew Sterenberg (14:50.701)

interviewing me as a student on the phone, let’s say, or through Zoom and my camera’s off. Like I could program my voice to answer you. there, it sounds crazy, but we are not that far away from being able to take my voice and answer questions that you have for me live, right? It’s not that far away really.

Emily Pacheco (15:11.502)

Yeah, right now it’s at the level of being able to write a dissertation and it started as a babbling toddler. And that wasn’t that long ago. That was November 2022 and here we are, 2025 and it’s only improving 100%. And I do wanna go back to something you said about doing away with the essay. I think we do need to question our assessments. As educators and that’s something that’s just not happening fast enough for sure. But you know, is this the best mechanism to get what we need out of applicants? And I’m not 100 % sure that what we’re asking for in the admission process is actually doing the job that we want it to do in the best way possible. So I think there is a lot of hope that this technology could actually, you know, create a new way that we could evaluate applicants and be able to kind of better figure out if they’d be successful at our institutions and evaluate their fit, not just based on the seven items that we are requesting at this point, which are the same seven items that we asked for in 1972 when we first asked for applications. And I don’t think that’s serving its purpose. It’s just not serving its purpose.

Matthew Sterenberg (16:28.589)

Yeah. And if you talk to, you know, different institutions, they’re going to give you different answers on what they find valuable. Right. I mean, everything has its problems, right. As a former admissions professional myself, the essay, I always felt like, it’s kind of a nice to have it. But a lot of times you either get some student that’s just bragging about themselves and how great they are, or it’s a sob story about how I know my grades aren’t good, but here’s why, you know, and you’re always like, Yeah, I don’t know if this is the, I don’t know if we needed this and some people will definitely disagree with that. And then other people say the GPA is not really good at accurate predictor of their success. Test scores are problematic. So how do we get a more holistic view? And I think that is part of the exciting part of AI that we can maybe rethink the information that we can input, how we can actually predict who’s going to be a successful student here or. the right fit for our institution. So I think a lot of that is exciting. You have a great website for this. I do want to plug that edhub.ai. You have a number of webinars that you’re doing on this topic. So if people are listening to this and they want to continue to engage with this topic, definitely check out edhub.ai. And you’ve got a number of different resources on there. And you really do a good job of addressing all the stakeholders that are part of this conversation. And it’s college counselors, it’s higher ed admissions professionals, it’s ed tech providers that are building technology or something that institutions will leverage. I want to talk about the college counselors, right? The people that are advising, like you told a little bit of a story of meeting with them, but when they sit down with students, what advice do you have for them? Because there is a lack of guidance from higher ed on how to do this. What should college counselors, how should they address this topic with their students?

Emily Pacheco (18:32.908)

Yeah, I think it’s a really important question because as I mentioned earlier, we know students are using this. They were the first ones to jump in and start experimenting with these tools. for a college counselor, not so much for a college counselor who just started last year or two years ago, three years ago. From my experience, those college counselors are pretty excited about these tools because they’re finding this is really helping them kind of make up for that lack of 10, 15 years of experience. What you find is when I sit down with a college counselor who has been in their role for five plus, 10 plus years, it can be really hard for them to see the usefulness of them joining this conversation. They say, I have my system. The way that I have been working with students, the way that I’ve been doing my college counseling, it works really well and I just don’t think I need these tools. And I know that’s a dangerous, you know, it’s a dangerous place to be because your students are using it and you are not using it and therefore you’re not engaging in conversations around that usage and figuring out kind of what that means for the way that you have been doing your college counseling because it does have to change now. And that’s a hard realization for many counselors. They want to believe that they can just kind of continue doing counseling in the way that they’ve been doing it. And that really is not the case because the way that students are approaching that admission process is very different now that they’re using these tools. I’ll give you a…

Matthew Sterenberg (20:09.987)

And if there’s a void of information on how to use it, what am I going to use it for? I’m going to use it for like, not to say the word cheat, but if no one’s telling me proper uses or helpful ways to use it, I’m just going to use it in the way that I feel benefits me the most instead of like thinking about how I’m using it. I’m not, if I’m not getting that instruction from a college counselor, I’m it’s, I’m just going to use it. However, I, I deem the best.

Emily Pacheco (20:33.454)

That’s right. Yeah. Yeah, and if you can’t sit next to an educator and say, look at this use, you know, what do you think about it? How do you even discover if that’s ethical or not? And I think a lot of them are actually doing pretty cool things that we would look at and go, wow, actually, I think that is ethical. I had never thought of that way to use this tool. And there’s other things, of course, that they’re like, look at this cool thing I’m doing, and that we’d look at and go, no, no, no, you know, let me explain to you why that isn’t the right way to use these tools in a learning environment. I recently sat with a college counselor. He’s at least 15, if not 20 years in his role at a pretty prestigious high school in Seattle. And he said he had sat with a student, he read the essay from the student, he knew the student really well. So he knew the way that the student wrote. And reading this essay, he said, I felt like he had used artificial intelligence in it, just knowing the student, I didn’t feel that this was something that they had created, but it was also what he would consider a five-star essay. And he said, it hit all of the marks. He said, there wasn’t anything that I could turn and say, hey, this is obviously created by AI because it’s not a good product. And he said, I didn’t know what to do. He was silenced. He proceeded with pretending that AI wasn’t in the room. And he told the student, this is incredible. I really think this is gonna be a successful college essay. And when I came that had just happened and this, know, and talking about my work in AI in admissions, he brought this story up and he said, I just didn’t have any words to discuss artificial intelligence with a student.

Matthew Sterenberg (22:21.593)

And you don’t want to levy accusations, but at the same point, I mean, you have to be really careful. Yeah.

Emily Pacheco (22:24.973)

Yeah. Yeah, and it was a good final product. mean, I’ve heard this story before, know, time and time before, time and time again before, but usually it’s where a counselor tells me, it was such a terrible essay. And they kind of, you know, making the point like, artificial intelligence is no good. You know, saw a student used it. It’s always so obvious to me, you know, when a student uses it. And I don’t think that’s the case. I think that’s the case when a student is new at using these tools. doesn’t have experience and hasn’t looked at other good quality essays and hasn’t really spent time. But when you take a student who’s savvy at technology, they’re going to learn how to make this look like a good essay.

Matthew Sterenberg (23:06.755)

Yeah, they literally could put in the prompt, like write it like a 17 year old would write it to make it sound not quite as good or so. Like there’s plenty of ways that you could, if you’re savvy enough to tone it down, honestly.

Emily Pacheco (23:24.502)

Yeah, and that’s what they’re doing. know, mean, that’s, you know, we’ve all been students, you know how to get things around that, you know, around the detection system if you want to. But I also think it’s important for us to remember that students aren’t just cheaters, you know, they don’t typically they don’t really, mean, know, studies show they don’t want to cheat. 5 % of students will cheat and AI doesn’t change that. And that’s really what we’re finding. This isn’t making a classroom of cheaters.

Matthew Sterenberg (23:27.098)

Yeah.

Emily Pacheco (23:53.582)

Students are cheating because they don’t know, like we’re mentioning, they don’t know the appropriate ways to use it. But I think with a little bit of guidance, they would just be using this in a way to enhance the learning process. And I think that’s the case in college admissions as well. If we want to teach them how to write a good essay, if we still deem that’s important, we can help it use this technology for certain aspects of that process and still teach that valuable writing process to them.

Matthew Sterenberg (24:27.523)

Yeah, I like the way you framed it in terms of enhancing, because I think that’s like AI is not good or bad. It is all about how the use of it. And then I think we need to get away from there’s a binary of did you use AI or didn’t you, or don’t use AI or use AI because there’s flavors of, you know, it’s a gray area. Like I used it for part of it, or I, I it’s you know, spit something out and then I edited that like it’s not, I use AI or did not, there’s all different versions of it. But then how does this enhance the experience? Like, I smarter for having used it or am I using it for something I don’t understand? Right? Like I could have it spit out a dissertation of which I know nothing about. Right. And that’s not making me smarter, but how do I use it in a way that’s actually enhancing my experience? How do I use it for it to teach me something? that’s a really, mean, I don’t know where we even begin there, but again, you could give me a prompt and I could ask it to spit something out, I give it back to you. That doesn’t mean I understood it at all. Or I could actually use it to help me understand something that then I could produce. So it’s such a difficult spot. But I do think it’s like, this making me dumber or smarter? It’s like the basic way I would put it. And there are ways to, like, that’s good writing. I see good writing. That’s a really clear way to frame that. I should use that mechanism the next time I write an essay instead of just relying on it solely.

Emily Pacheco (25:57.954)

Yeah. Yeah, yeah, and I hear you saying that from a perspective of a student, but I think you can even take that from a perspective of a college admissions counselor, you know, and say, how can I use this to enhance my work? You know, and I think it’s kind of making that change from viewing this as Google, as I see it, you know, like where I can use it to put in a question and get an answer, but instead, how do I enhance myself as a professional? You know, this is what I’m able to do as a college counselor. These, you know, I can service students in this way, this number of students, but how can I provide, you know, more personalized attention or do something in a different creative way using these new tools? And that, you know, that takes a switch in our brain. You know, we’re not programmed to use AI at all. Our brains, is all new. We’re very programmed to use Google and we try to use AI like Google, you know, and that’s where a lot of educators, I think,

Matthew Sterenberg (27:03.215)

Mm.

Emily Pacheco (27:05.122)

run into a wall of being like, these tools are terrible. I put in a prompt and what I got back was awful. And when you start digging into it, you realize pretty quickly it’s because we don’t yet understand how to appropriately prompt artificial intelligence. It’s a very different process. a, you know, and it takes a different way of thinking. But when I’ve seen that kind of switch go off with college counselors where they understand this can actually just make them a better college counselor. It’s not looking to replace them. You know, I think almost every parent, every student would love to have a human college counselor to work with. But the reality of the situation is there’s just not enough quality human counselors for every student. So I really think this technology could do a lot to provide quality counseling to more students.

Matthew Sterenberg (27:55.447)

Emily, is there anything we didn’t highlight in our conversation that you want to make sure we cover for the folks listening? Anything that you want to make sure people know about AI and college admissions before we wrap?

Emily Pacheco (28:09.964)

Yeah, I think it’s important that every office, so every admission office, every college counseling office starts to think about who in their office is using this technology and bring them to the table to kind of more systematically have these conversations about what does it mean for our office? What does it mean for our students, you know, who are using this technology? Because people in your office are using it. I always say go to the newest employee that’s most likely the person who is using it in the most creative ways and make sure that your office is a place where people can talk about this because if these conversations can’t take place, then we can’t really move things forward. So if you’re not sure where to start, if you are an educator, definitely join the edhub.ai. We are a special interest group recognized under NACAC. So we’re really active as well within the NACAC organization and doing a lot with them as well. And yes, we’ve got a webinar later this month on the 28th of February that’s going to be looking at the ethical concerns about this. We always have those questions come up. So we’ve dedicated an hour and bringing on a panel of experts to kind of talk more deeply about some of those ethical questions specifically around using AI in that admissions process. Great place to start, join us in that conversation, but make sure to join our group on LinkedIn as well.

Matthew Sterenberg (29:35.109)

All right, so check out edhub.ai. If you’re a higher ed institution, you’re probably coming up with guidance for your students on how they use AI. Come up with some guidance for how your prospective students should use AI. That’s a big takeaway for me. If you’re a college counselor and you’re having a college night, you’re bringing in parents and students, talk about AI. Talk about how you expect them to use it, how they can leverage it. Some great takeaways. Emily. Thank you so much for joining me.

Emily Pacheco (30:05.966)

Thank you, Matt. This has been a lot of fun.

There’s always more to learn.